Simulation of flow and Sedimcnf Transport in an open Channel with

Obstacle using IRIC NAYSZDH

M*ruma H*que

Ilepartment of Civil Enginecring

Khulnn Khulna University of Engineering and Technology
Khulna-9203n Bangladesh
M*yr 2018

i
Simulation of {Iow and Sediment Transport in an Open Channel with
Obstacle using IRIC NAYS2DH

A Project Report Submitted to the Department of Civil Engineering, Khulna Khulna

University of Engineering and Technology in Partial Fultillment of the Requirements tbr the
Degree of

'oMasters of Seience ir Civil Engineeringo'

Supervised by: Prepared by

Dr. Md. Shahjahan AIi Masuma Haque

Professor Roll No: 1301509

Department of Civil Engineering I)epartment of Civil Engineering

KUET, Khulna-9203 KUET, Khulna-9203

Department of Civil Engineering

Khulna Khulna Univcrsity of Engineering and Technology
Khulna-9203, Bangladesh
M*yr 2018
 Ileclaration

This is to certifu that the thesis work entitled " Simulation rf Flarv and Sediment Transport in
an Open channel rvith Obstacie using iRiC NAYS?Di{" has been carried our by Masuma
Haque in Department of Civil Engineering, Khulna University of Engineering & Technology,
Khulna, Bangladesh. The abor,e thesis rvork or any part af this rvork has not been submitted
anvrvhere fbr the an"ard of any'degree or diploma.

lvlast@-
Dr. Md. Shahjahan Ali ldasuma Haque

(Superv'isor) RollNo: 1301509

Prolessor

Department of Civil Engineering

KUET, Khulna-9203

iii
 Approval
This is to certify that the thesis work submitted by Masuma Haque entitled 'Simulution of Flow
and Sediment Transport in an Open Channel with Obstacle using iRIC Noys2DH'has been
approved by the board of examiners for the partial fulfillment of the requirements for the degree
of Master of Science in Civil Engineering in the Department of Civil Engineering, Khulna
University of Engineering & Technology, Khulna, Bangladesh in May 29,2018.

BOARD OF EXAMINERS

1. Chairman
(Supervisor)
Dr. Md. Shahiahan Ali
Professor, Department of Civil Engineering
Khulna University of Engineering & Technology

2. Member

Dr. Md. Shahjahan Ali

Head, Department of Civil Engineering
Khulna pniversity of Engineering &Technology

-). N\, Member

Dr. Kh. Md. Shafiul Islam

Professor, Department of Civil Engineering
Khulna University of Engineering &Technology

4. -.4*--ir, ;fryc/g Member

Dr. Khondoker Mahbub Hassan

Professor, Department of Civil Engineering
Khulna University of Engineering & Technology

). Member
Dr. Md. Munsur Rahman (External)
Professor, Institute of Water and Flood Management
Bangladesh University of Engineering & Technology
ACKNOWLEDGEMENT

All praises go to almighty Allah, the most magnificent merciful. The author gratefully
acknowledge their profound gratitude and indebtedness to project supervisor, prof. Dr. Md.
Shahjahan Ali, Assistant professor, Department of Civil Engineering, Khulna University of
Engineering& Technology (KUET) for his keen interest in this project, continuous
supervision, guidance, encouragement, inspiration and thoughtful suggestion in completing
the work.

Finally author would like to express her gratitude to all the faculty members of the
Department for their kind Co-operation.
ABSTRACT
Hydraulic structures such as barbs, groins and spur dikes have been constructed in river bank
in order to deflect the flowing water away from vulnerable zone. They are made of stone,
gravel, rock, earth or piles, beginning at the river bank with a root and ending the regulation
line with a head. River flow mechanisms with the hydraulic structures need to be intensively
studied to protect the river bank from erosion, increase the bank stability, improve navigation
and flood control. Such structure also plays an important role to enrich the biodiversity of
different aquatic species by providing shelter them. The main objective of this thesis is to
simulate the straight channel with the presence of several types of such structures by using
2D numerical simulation model, iRIC Nays2DH. The groin like structures can be oriented
perpendicular or be inclined either upstream or downstream. Each orientation affects the
stream in a different way and shows different result in the vicinity of the groin. In chapter,
simulation of flow around groin, details of streamline variation, velocity vector and computed
recirculation around groin are shown for 45º, 90º and 135º groin. The length of recirculation
zone is found 1.2m, 1.04m and 1m, respectively. The simulated present study of velocity
profiles and shear stress profiles are compared with previous experimental and numerical
results. Good agreement is found for all of the cases. It is observed that predicted velocity
profiles and predicted bed shear stress profiles for 45º, 90º and 135º, the position of
maximum velocity and shear stress are found to be shifted towards downstream with
increasing y/l = 1, 1.5, 2, 3and 4, respectively. For all the cases, peak of the velocity and
shear stress are also deviated from upstream to downstream, 45º to 135º angled groin. From
the velocity contour and shear stress contour, it is found that left bank is higher velocity and
higher shear stress zone and right bank is lower velocity and lower shear stress zone. In
chapter, local scour around groin, to determine the variation of scour depth and sediment
deposition near the head of the groin, groins are orientated 45º, 90º and 135º at the upstream
of the channel. The comparison of velocity field, bed level contour results, Elevation of bed
profiles of 45º, 90º and 135º groin are given. Simulated present study are compared with
previous numerical results and good agreement are found. 45º, 90º and 135º groin are
simulated for 1 hour and 5 hours, respectively. After 5 hours, equilibrium scour is attended
and maximum scour depth for 45º, 90º and 135º are found as 0.065m, 0.062m and 0.054m,
respectively. In case of multiple groin, maximum scour are found 0.145m for discharge
0.035 m3/s and 0.24m for discharge 0.046 m3/s.. Bed deformation contour lines, velocity
vectors, transverse and longitudinal bed profiles are depict in this chapter. Computed
scouring are compared with previous experimental and numerical results. For discharge,
0.035 m3/s, scour at first spur dikes to third spur dikes are 0.123m, 0.005m and 0.005m,
respectively. Similarly, when discharge 0.046 m3/s, scour at first to third spur dikes are
0.16m, 0.01m and 0.012m, respectively. The model is also used to evaluate and analyze flow
patterns, streamline variation, velocity vector contour, vorticity, bed shear stress contour,
scour and sediment deposition around a groin like structure on a straight channel.

Most of the natural flow are three dimensional; for example, in case of meandering channel,
the flow passing through the bend is obviously of a three-dimensional (3D) nature because of
the generation of secondary current due to the centrifugal force. When a flow in an open
channel passes an obstacle ( such as groin), the deviation of flow causes the generation of
secondary current and flow become highly 3D. Therefore, a 3D hydrodynamic model is
necessary to accurately simulate the flows in a meandering channel. However, dealing with
the practical engineering problems, such as alluvial geomorphic process, it is not
computationally efficient to use 3D models. In such problems, a two-dimensional (2D) model
is generally used. To predict the applicability of iRIC software in a flow field with strong
secondary current, the model is also applied for 180º sharp and mildly bend. Water surface
elevation contour and velocity vector are evaluated from the simulation. For both cases, the
velocity of flow is greater in the inner section than the outer section. The results are
compared with previous experiment result and found well agreed with previous experiments.
Table of Contents
ABSTRACT .......................................................................................................................................i
ACKNOWLEDGEMENT .................................................................................................................ii
TABLE OF CONTENTS ...................................................................................................................iii
LIST OF TABLES .............................................................................................................................vi
LIST OF FIGURES ...........................................................................................................................vii
Chapter 1 INTRODUCTION .........................................................................................................1
1.1 GENERAL ............................................................................................................................1
1.2 OBJECTIVE OF THE STUDY .............................................................................................3
1.3 ORGANIZATION OF THE REPORT ..................................................................................3
Chapter 2 LITERATURE REVIEW ............................................................................................4
2.1 INTRODUCTION ...................................................................................................................4
2.2 REVIEW OF PREVIOUS STUDIES ......................................................................................4
2.3 GENERALIZED SIMULATION PROCEDURE ...................................................................5
2.3.1 Simulation and modeling .................................................................................................5
2.3.2 Summary of simulation process ......................................................................................5
2.3.3 Components of a numerical solution method ..................................................................6
2.3.4 Discretization approaches ................................................................................................6
2.3.4.1 Finite Difference Method .............................................................................................7
2.3.5 Numerical grid .................................................................................................................8
2.3.6 Properties of numerical simulation methods.....................................................................8
2.3.7 Courant number ................................................................................................................11
2.4 iRIC SOFTWARE ..................................................................................................................11
2.4.1 Features of the flow field calculation model ................................................................................ 12
2.4.2 Features of the sediment transport and riverbed deformation calculation ........................13
2.4.3 other feathers .....................................................................................................................14
2.5 BASIC EQUATIONS ..............................................................................................................14
2.5.1 Basic flow equations .........................................................................................................14
2.5.1.1 Basic equations in orthogonal coordinates.....................................................................14
2.5.1.2 Transformation into general curvilinear coordinates .....................................................15
2.5.1.3 Basic equations in a general curvilinear coordinate system ..........................................16
 2.5.2 Turbulent field calculation method ...................................................................................18
2.5.2.1 Constant eddy viscosity .................................................................................................18
2.5.2.2 Zero-equation model ......................................................................................................18
2.5.2.3 K-ε model ......................................................................................................................19
2.5.3 Bottom friction calculation method .................................................................................19
2.5.3.1 Constant value ................................................................................................................19
2.5.4 Method of calculating resistance by vagitation.................................................................20
Chapter 3 SIMULATION TECHNIQUES ...................................................................................21
3.1 INTRODUCTION ...................................................................................................................21
3.2 SIMULATION TECHNIQUES ..............................................................................................21
3.3 NAYS2DH ...............................................................................................................................23
3.4 OPERATIONAL PROCEDURES ..........................................................................................23
3.5 FLOW PARAMETERS ...........................................................................................................23
3.6 COMPUTATIONAL SCHEMES AND FLOW DOMAIN ....................................................25
Chapter 4 SIMULSTION OF FLOW AT A BEND ....................................................................33
4.1 Introduction ..............................................................................................................................33
4.2 Simulated result of 180º sharply curved channel ....................................................................35
4.3 Simulated result of Mildly curved channel .............................................................................38
Chapter 5 SIMULATION OF FLOW FIELD AROUND GROIN LIKE STRUCTURE ........41
5.1 INTRODUCTION ...................................................................................................................41
5.2 NUMERICAL TEST ...............................................................................................................41
5.2.1. Flow domain and hydraulic parameter.............................................................................41
5.2.2 Computational Scheme .....................................................................................................43
5.3 SIMULATED FLOW FIELD ..................................................................................................43
5.3.1 Streamline .........................................................................................................................44
5.4 COMPARISON OF SIMULATION RESULS WITH PREVIOUS STUDIES ......................45
5.4.1 Computed velocity profiles ...............................................................................................45
5.4.2 Computed bed shear stress ................................................................................................47
5.5 SIMULATED RESULTS FOR DIFFERENT CASE .............................................................49
5.5.1 Predicted velocity profiles ................................................................................................49
5.5.2 Predicted bed shear stress profiles ....................................................................................52
5.6 COMPUTED VELOCITY CONTOUR ..................................................................................54
 5.7 COMPUTED SHEAR STRESS ..............................................................................................56
Chapter 6 LOCAL SCOUR AROUND A GROIN .......................................................................58
6.1 INTRODUCTION ...................................................................................................................58
6.2 BED DEFORMATION AROUND SINGLE GROIN ...........................................................58

6.2.1 Velocity field ....................................................................................................................59

6.2.2 Bed elevation around the structure....................................................................................64

6.2.3 Flow depth ........................................................................................................................75

6.3 SIMULATION OF CHANNEL WITH MULTIPLE GROIN ................................................78

6.3.1 Flow Domain and Hydraulic Parameters ..........................................................................78

6.3.2 NUMERICAL SIMULATION RESULT .........................................................................79

Chapter 8 CONCLUSION AND RECOMMENDATION ...........................................................87

8.1 Conclusion ...............................................................................................................................87

8.2 Recommendation .....................................................................................................................88

References ........................................................................................................................................89
List of Figure

Figure 1.1: Schematic model of groyne-surrounding flow. (Sarveram et al., 2012) .........................2

Figure 2.1: Cartesian grid for FD method ..........................................................................................7

Figure 2.2: Finite difference grid .......................................................................................................8

Figure 2.3: Conceptual relationship between consistency, stability and convergence ......................9

Figure 2.4: An outline of the iRIC Software, its functions and features ...........................................12

Figure 3.1: Definition sketch and grid of an open channel with 1800 sharply curved bend ..............26
Figure 3.2: Definition sketch and grid of an open channel with 180o mildly curved bend ...............27
Figure 3.3 : Sketch and preparation of grid of a straight channel by iRIC Software .........................28
Figure 3.4: Sketch and barb installation for 45º angled to the flow direction ..................................28
Figure 3.5: Sketch and barb installation for 90º angled to the flow direction ..................................29
Figure 3.6: Sketch and barb installation for 135º angled to the flow direction ................................29
Figure 3.7 : Sketch and barb installation for double barb on both side .............................................30
Figure 3.8 : Sketch and grid preparation of a straight channel with three spur dikes ........................30
Figure 3.9: Groin orientation for 90º angled to the flow direction ....................................................31
Figure 3.10: Groin orientation for 45º angled to the flow direction ..................................................31

Figure 3.11: Groin orientation for 135º angled to the flow direction ................................................32

Figure 4.1: Definition sketch of the flow pattern at the straight portion of the river.........................34

Figure 4.2: Definition sketch of the flow pattern at the bend of the river .........................................34

Figure 4.3: Distribution of water surface elevation for Case II (in m), (a) Experiment by
Rozovskii (1961), (b) Simulated by iRIC Nays……………………….. ...........................................36

Figure 4.4 : Comparison of simulated water surface profile with experimental results for 180o
sharply curved channel ......................................................................................................................37

Figure 4.5: Comparison of width wise variation of velocity vector at different section of the
180o sharply curved channel (a) Experimental by Rozovskii (1961), (b) Computation by iRIC
Nays2DH............................................................................................................................................37

Figure 4.6: Velocity vector super imposed on the surface elevation contour (a) Experimental
(b) Simulation by iRIC Nays .............................................................................................................38

Figure 4.7: Distribution of depth-averaged velocity 180o mildly curved channel (a)
Experiment by De Vriend (1979), (b) Simulation by iRIC Nays2DH ..............................................39

Figure 4.8: Comparison of simulated water surface profile ..............................................................40

Figure 5.1: Computed velocity vector for (a) case-1, (b) case-2, (c) case-3 developed by iRIC
Nays2DH............................................................................................................................................42

Figure 5.2: Computed recirculation around groin for (a) case-1, (b) case-2, (c) case-3
developed by iRIC Nays2DH ............................................................................................................44

Figure 5.3: Streamline around a groin for (a) case-1, (b) case-2 and (c) case-3 developed by
iRIC Nays2DH ...................................................................................................................................44

Figure 5.4 : Comparison of resultant velocity profiles (W) with the available previous studies
for 90° groin ( case 1) ........................................................................................................................47

Figure 5.5: Comparison of bed shear stress profiles (τ) of predicted results with the available
previous studies ..................................................................................................................................49
Figure 5.6: Comparison between resultant velocity profiles ( W) of present studies for case-1,
case-2 & case-3 along lateral distance (a) y/l=1, (b) y/l=1.5, (c) y/l=2 ,(d) y/l=3, (e) y/l=4; ..........52

Figure 5.7 : Comparison between computed bed shear stress ( ) of the proposed model for
case 1, case 2, and case 3 along lateral distance, y/l = 1, 1.5, 2, 3, and 4 ..........................................54

Figure 5.8 : Velocity contoured around a groin for ( a) case -1 (b ) case-2, and case 3 at the
end of simulation by iRIC Nays2DH .................................................................................................55

Figure 5.9: Bed shear stress contour around a groin for test (a ) case 1 (b) case 2 and ( c) case
at the end of simulation by iRIC Nays2DH .......................................................................................56
Figure 6.1 : Velocity field simulation result at water surface into the channel without barb
after 1 hour simulation (a ) by Hossain et. al ( 2012) ( b )present study ...........................................60
Figure 6.2 : Velocity field simulation result at water surface into the channel without barb
after 5 hours simulation (a) by Hossain et. al (2012) (4) present study ..........................................61
Figure 6.3 : Velocity field with a barb installed one side of the channel after 1 hour
simulation (a)by Hossain et al (2012) (b) present study ....................................................................62
Figure 6.4: Velocity field with a 45° barb installed at one sidewall of the channel after 5 hour
simulaion ( a)Hossain et. al.,2012). ( 8 ) present study .....................................................................62

Figure 6.5 : Velocity field with 90° barb installed at one side of the channel (`9 ) after 1 hour
simulation. (10) after 5 hours simulation……………………………….. .........................................63

Figure 6.6 : Velocity field with 135° barb installed at one sidewall of the channel (11) after 1
hours simulation (12 ) after 5 hours simulation .................................................................................64
Figure 6.7 : Velocity field result with double barb installed at both sides at the channel (13)
after 1 hour simulation ( a) by Hossain et. al (2016) (b)present study ..............................................64
Figure 6.8 : Velocity field result with double barb installed at both sides at the channel (15)
computed simulation after 5 hours .................................................................................................................. 64

Figure 6.9 : Bed level contour result of the channel without any structure after 1 hour
simulation ( a )Hossain et. el., 2012), ( b ) present study .................................................................65
Figure 6. 10 : Bed level contour result of the channel without any structure after 5 hour
simulation ( a) by Hossain et. al ( 2012) ( b) present study ...............................................................66

Figure 6.11: Bed level contour result of the channel with a barb installed at sidewall of the
channel ( a ) by Hossain et. al (2012), (b ) present study ................................................................. 66
Figure 6.12 : Bed level contour result of the channel with a 90° barb installed at one sidewall
of the channel after 5 hour simulation ...............................................................................................67
Figure 6.13 : Bed level contour result of the channel with a 90° barb installed at one sidewall
of the channel after 5 hour simulation ...............................................................................................67
Figure 6.14: Bed level contour result of the channel with a 135° barb installed at one
sidewall of the channel (a) after 1 hour simulation, (b): after 5 hour simulation) .............................68
Figure 6.15: Bed level contour line after installed barb both side of the channel top: after 1
hour simulation ( a) by Hossain et. al., 2016), (b ) present study ......................................................68
Figure 6.16: Evolution of bed profile & Comparison before installing barb (a) after 1 hour
(b) after 5 hours ..................................................................................................................................69
Figure 6.17: Evolution of bed profile & Comparison (C) after 1 hour with 45º barb (d) after 5
hours with 45º barb……………………. ...........................................................................................70
Figure 6.18: Evolution of bed profile (e) & (f ) represents bed elevation with 90° barb after 1
hour and 5 hours……………………….............................................................................................71
Figure 6.19 : Evolution of bed profile & comparison ( g ) & ( h ) represent bed elevation with
135° after 1 hour and 5 hours, respectively .......................................................................................72
Figure 6.20 : Evaluation of bed profile and comparison of double at both side bed evaluation
(i) after 1 hour ....................................................................................................................................73

Figure 6.21: Evaluation of bed profile and comparison of double barb at both sides & bed
elevation of double barb (j) after 5 hours...........................................................................................73
Figure 6.22 : Comparison of 45°, 90° and 135° barbs (k) after 1 hour..............................................74
Figure 6.23 : Comparison of 45° , 90° and 135° barbs after 5 hour ..................................................74

Figure 6.24: Water depth contour result of the channel with 135° barb installed at one
sidewall of the channel (a) after 1 hours simulation, (b) after 5 hours simulation) ..........................75

Figure 6.25: Water depth contour result of the channel with a barb installed at one sidewall of
the channel after 5 hours ( a) by Hossain et. al.,(2012) (b ) present study ........................................76

Figure 6.26: Water depth contour result of the channel with a 45° barb installed at one
sidewall of the channel after 1 hour simulation (a) by Hossain et. al (2012), (b) present study .......76

Figure 6.27: Water depth contour result of the channel with a 45° barb installed at one
sidewall of the channel after 5 hour simulation ( a) by Hossain et. al.,(2012) (b) present study ......77
Figure 6.28: Water depth contour result of the channel with 90° barb installed at one sidewall
of the channel ( a) after 1 hours simulation, (b) after 5 hours simulation) .......................................77

Figure 6.29: Water depth contour result of the channel with 135° barb installed at one
sidewall of the channel (a)after 1 hours simulation, (b) after 5 hours simulation) ...........................78

Figure 6.30: Bed deformation contour line of case S4C1 simulated by iRIC Nays2DH ....................80

Figure 6.31: Bed deformation contour line of test S4C2 simulated by iRIC Nays2DH .....................80

Figure 6.32 : Computed velocity vector in case of test S4C1 developed by iRIC Nays2DH ...........81

Figure 6.33: Computed Velocity vector in case of test S4C2 developed by iRIC Nays2DH ............81

Figure 6.34: Definition sketch for different cross and longitudinal sections, ....................................81

Figure 6.35 Bed changes comparison of the E1 results found by iRIC Nays2DH with
FLOW-3D and experimental results ..................................................................................................83

Figure 6.36: Bed changes comparison of the E2 results found by iRIC Nays2DH with
FLOW-3D and experimental results ..................................................................................................85
List of Table
Table 2.1 Model constants .................................................................................................................19

Table 3.1 Hydraulic parameters for numerical simulation in series 1 ..............................................24

Table 3.2 Hydraulic parameters for numerical simulation in series 2 ..............................................24

Table 3.3 Hydraulic parameter for numerical simulation in series 3.................................................25

Table 3.4: Hydraulic parameters for numerical simulation in series 4 ........................................................... 25

Table 6.3: Comparison of the scouring amount in iRIC Nays2DH results with Flow-3D
(pourshahbaz et al., 2017) and experimental results (karami et al., 2014) .......................................85
List of Symbols
ρ Density

u Velocity in the x direction

v Velocity in the y direction

w Velocity in z direction

ΔA Surface area

Σ Sum of the external forces on the control volume

t Time

Δt Time step

h Water depth

g Gravitational acceleration

H Water depth

τx Riverbed shearing force in the x direction

τy Riverbed shearing force in the y direction

Fx Resistance by vegetation in the x direction

Fy Resistance by vegetation in the y direction

Cf Riverbed shear coefficient.

νt Eddy viscosity coefficient.

CD Drag coefficient of vegetation.

as Area of interception by vegetation per unit volume.

θ Represents the angle formed by the x axis and the ξ axis (or the y
axis and the η axis).

vt Eddy viscosity coefficient

l Length

α Proportional constant

k Karman coefficient

Cμ Model constant
C1ε, C2ε, σk and σε K-ε model constants

Cf Coefficient of riverbed shearing force

n Manning's roughness parameter

ks Relative roughness height

d Sediment diameter

g Gravitational acceleration.

Cd Drag coefficient of vegetation

αs Vegetation per unit volume

ns Number of vegetation

Ds Average diameter of trunks

Ss Sampling grid width.

L Length of the straight portion

Qo Upstream discharge

ho Mean depth

S Bottom slope

n Manning’s coefficient

Vr Velocity ratio

uo Mean velocity
 Chapter 1
INTERODUCTION
1.1 GENERAL

In the field of hydraulic engineering and river engineering the most commonly used
structures are the groins, spur dykes and barbs. These hydraulic structures have been
constructed in river bank in order to deflect the flowing water away from vulnerable zone.
For the purpose of river training, construction of such obstacles against water flow causes
significant change in flow patterns, sediment transport and bed topography. The main
purpose of built up of such obstacles on natural river bank is to divert the direction of water
flow so that bank erosion can be eliminated. The shape of obstacle is similar to a trapezoidal
structure with inclined sides and a wide sloped crest. This obstacle behaves as a partially
submerged structure (weir) when flow is low and fully submerged when bankfull flow
conditions are present. When pointed upstream, the submerged weir section forces the water
flowing over the structure into a hydraulic jump (Fox, 2002). The flow separation induced by
the hydraulic jump promotes the formation of eddies and sediment deposition on the leeward
side of the barb (LIoyd and Stansby, 1997). Generally, barbs are used to protect banks for
gentle (wide radius) menders, or relatively straight banks. The primary function of barbs is to
deflect the strong main flow away from critical zones towards the channel centre and
therefore, prevent erosion of banks (Hansen and Winter, 1996). The approaches to protect
stream banks from erosion during periods of high flows including various bio-technical type
channel revetments (Schiechtl and stern, 1997) and groin like structure including bend-way
weirs (Davinroy,1990,1994) and barbs (USDA, 2001). Barbs have been used by the Natural
Resources conservation Service (NRCS) of US department of agriculture, in Oregon for river
and stream bank protection since the late 1980. According to Chow (1959), barbs are a
trapezoidal shaped rock structure, which extends into the main flow of the river.

Many experimental and numerical researches have been done in order to examine flow
pattern and scouring around groins (Rajaratnam and Nwachukwu, 1983). Those studies were
carried out in different conditions in terms of groyne length, groyne installation angle towards
the approaching flow, permeable or impermeable states, submerged and non-submerged
states and number of groins, etc. (Yeo et al., 2005). Due to the variety of groins flow
separation and recirculating length would be greatly different which is a challenge to the
applications of numerical models (Quanhong and Pengzhi, 2007). The groin which is located
at the river main flow with perpendicular orientation, causes narrowing and deviation of the
flow and this in turn gives a complex three dimensional form to flow around a groin
(Tingsanchali and Maheswaran, 1990). As local scouring, sediment transport and settlement
around groin require clarifying the two-dimensional flow pattern, it is necessary to consider
the flow pattern around a groin, which is important and necessary to conduct experimental
examinations (Jamieson et al., 2013). However, for a full understanding of this flow pattern,
one has to use numerical techniques.

1
 Figure 1.1: Schematic model of groyne-surrounding flow. (Sarveram et al., 2012)

In the flow field with groyne or spur dikes, the flow is separated at upstream of spur dikes
and went to levees and created vortexes move to the downstream which finally causes local
deposition surface scouring around the structure. In the usage of spur dikes for any purposes,
maximum scour depth and scour around spur dikes are so important as can be seen in many
researches such as: Ahmad (1951,1953), Garde et al (1961), Gill (1972), Richardson et al
(1975), Rajaratnam et al (1983), Shields et al (1995), Kuhnle et al (1999), Kothyari et al
(2001), Barbhuiya et al (2004), Kang et al (2011). These structures are mainly built in groups
and the flow characteristic would vary in peak and between spur dikes. Researchers such as
Chang et al (2013) and Karami et al (2014) investigated the scour depth in groups of spur
dikes. Due to the greater interaction between flow and sediment and revealing turbulent flow
in group of spur dikes, exact scour depth forecasting needs consideration of flow turbulence
in scouring hole Mendoza (1993).

Most of the natural flow are three dimensional for example, in case of meandering channel,
the flow passing through the bend is obviously of a three-dimensional ( 3D) nature because of
the generation of secondary current due to the centrifugal force. When a flow in an open
channel passes an obstacle ( such as groin), the deviation of flow causes the generation of
secondary current and flow become highly 3D. Therefore, a 3D hydrodynamic model is
necessary to accurately simulate the flows in a meandering channel. However, dealing with
the practical engineering problems, such as alluvial geomorphic process, it is not
computationally efficient to use 3D models. In such problems, a two-dimensional (2D) model
is generally used. To predict the applicability of iRIC software, the model is also applied for
sharp 180º and 180º mildly curved bend to simulate the flow field.

2
In this study, several case studies for open channel flow are simulated after installing the
groins with 45°, 90° and 130° angled with the river bank. In addition to the flow field, the
local scour around single and multiple groins are also studied.

1.2 OBJECTIVE OF THE STUDY

The main objective of this study is to develop numerical model using iRIC NAYS2D (a
computational flow model for multi-dimensional river flow and river bed variation analysis
solver package) and to examine the flow and sediment transport behavior in open channel
with different topographic configuration and obstacles. The specific objective of this study
are outlined as below:
 To simulate two-dimensional open channel flows with 180° sharply curved and
mildly curve channel to examine the performance of the software in determining the
flow field with strong secondary current.
 To simulate the flow field around a groin with different orientation and flow
conditions.
 To simulate the local scour around single and multiple groin fields/stream barbs.

1.3 ORGANIZATION OF THIS REPORT

This thesis is divided into seven chapters. In Chapter 1, the introduction is mentioned. In
Chapter 2, Review of literature and detail descriptions of basic equations used for simulation
are given. The simulation technique is illustrated in Chapter 3. In Chapter 4, Simulation of
flow at a bend is discussed, in which the secondary flow behavior in a sharply and mildly
bend channel is studied by using iRIC Nays2DH. In Chapter 5 discussed about simulation of
flow field around groin like structure where performance of 45º, 90º and 135º angled groin on
the flow direction of an open channel is studied. In Chapter 6, the detail of sediment transport
around groin is mentioned, where scour and deposition phenomena is studied in the presence
of different angle barb on the flow of a channel. Besides, multiple spur dikes performances
are studied in this chapter. In addition, the Chapter 7 contains the conclusion of the present
study and recommendations for further study.

3
 Chapter 2

LITERATURE REVIEW

2.1 INTRODUCTION

In this chapter, the available previous studies on experimental and numerical results are
reported on groin/stream barb/spur dikes. Besides, some basic equation and concept of
iRIC Nays2DH software are illustrated

2.2 REVIEW ON PREVIOUS STUDIES

Kim and Choi (2003) conducted an experiment on numerical simulations of open-channel

flows in a bend using the finite element method. They used the 2D numerical algorithm for
this simulation. The proposed model was applied to 180° bend experimented by Rozovskii
(1961). The simulated results compare favorably to measured data. Then, the model was used
to simulate flows in a 7.7 km curved reach in the Han River, Seoul, Korea. They also
investigate the impact of planting vegetation on each side of floodplain.

Hsieh and Yang (2003) studied the suitability of 2D models for bend flow simulation by
using both a conventional model and a bend-flow model. Analysis of the simulation results
indicated that the maximum relative difference in longitudinal velocity is mainly related to
the relative strength of the secondary current and the relative length of the channel. Empirical
relations connecting the maximum relative difference in longitudinal velocities, the relative
strength of the secondary current, and the relative length of the channel, were all proposed to
be used as a guideline for model users.

Hossain et al. (2012) applied a numerical model on a rectangular straight channel which is
13.3m long and 0.8m wide. The barb structure is installed as a obstacle at 45°, 90° and 135°
on the upstream portion of river channel. The length of the barb is 24cm, one third of the
cross-section. The numerical simulations have done considering sediment transport. Barbs
have been used by Natural Resources conservation Service (NRCS) of US department of
agriculture, in origon for river and stream bank protection since the late 1980. The
Washington State Department of Transportation (WSDOT) has employed barbs for bank
protection along highways or river crossing and /or to improve aquatic conditions, especially
in shallow gravel bed streams (Papanicolaou et al. 2004). There are various researches have
done for flows around hydraulic structures like spur dikes, barbs, groins, submerged vanes
etc, the majority of the investigations have treated flow around these structures and prevent
erosion of the bank. Kothyari andRaju (2001) reported that, the primary focus of previous
investigations, such as, Gill (1972), Grade et al. (1961) and Ahmad ( 1953), on the study of
scour at abutments and spur dikes was to predict an equilibrium or maximum design scour
depth. Barbs induced energy re-distribution away from the outer bank towards the center of
the channel results in scour near the ends of the barbs and realignment of the thalweg. Field
4
observation and laboratory result showed by Johnson, et. al. (2001), Matsuura and Townsend
(2004), Kuhnle et. Al.( 2002) that, the scour depth occurs at the barbs end and immediately
downstream of the structure.

The research in flow field carried out by Francis et al. (1968) was considering separation
zone for various types of groynes in a rectangular flume, but they did not measure the flow
velocities. Researches done by Rajaratnam and Nwachukwu (1983) contain velocity
measurements in the flow field in the presence of groin. Ettema and Muste (2004) examined
the effect of groyne length on downstream separation region by considering the effect of
scale. Researches by Yeo et al. (2005) also deal with an experimental examination of the
downstream separation region of a groin under various groyne lengths, various installation
angles and various degrees of groyne permeability (Zhanfeng and Xiaofeng, 2006).

Tingsanchali and Maheswaran (1990) used a two-dimensional depth-averaged model,

incorporating a correction factor used in the k–ε model and introducing a three-dimensional
correction factor to improve the computed bottom stresses. Molls et al. (1995) also developed
a general mathematic model to solve the unsteady two-dimensional depth averaged equations
by combining it with a constant eddy viscosity turbulent model (Ettema and Muste, 2004). A
great deal of research has been carried out in recent years, especially by Chinese researchers
such as Zhanfeng and Xiaofeng (2006), Quanhong and Pengzhi (2007) and Tang and Ding
(2007). They studied the model and scouring around groyne by using various models of
Turbulence under different conditions of flow and groyne dimension.

2.3 GENERALIZED SIMULATION PROCEDURE

2.3.1 Simulation and modeling
Simulation of a system is used for understanding the physics of the process of the system in
time and space. For example in water resources system: effects of land use and climate
changes on the balance, groundwater levels, stream flow variability, water quality, etc.

• In the present context the term `model' refers to simulation model and such a model is
a representation of natural or human constructed world.

• This model is simpler than the prototype system and can reproduce some
characteristics thereof.

• The simulation model produces series of outputs.

2.3.2 Summary of simulation process

Generally the simulation process contains following steps:

• Conceptualization of process of physical system.

5
 • Transforming the process of physical system by into partial differential equation.

• Transforming the partial differential equation into algebraic equations or numerical

equations.

• Development of the solution algorithm of the numerical equations.

• The computed codes are written.

• Finally, the model is validated.

2.3.3 Components of a numerical solution method

A numerical solution method contains following components –

• Mathematical Model.

• Discretization Approach.

• Coordinate and Basis Vector Systems.

• Numerical Grid.

• Finite Approximations.

• Solution Method.

• Convergence Criteria

Discretization Approach and Numerical Grid are two most important components. Here we
will explain these.

2.3.4 Discretization approaches

There are three discretization approaches-

• Finite Difference Method

• Finite Volume Method

• Finite Element Method

In the present code that has been used here is Finite Difference Method. So, this method is
explained below in brief.

6
2.3.4.1 Finite Difference Method
It is the easiest method to use for simple geometries. The starting point is the conservation
equation in differential form. The solution domain is covered by a grid. At each grid point,
the differential equation is approximated by replacing the partial derivatives by
approximations in terms of the nodal values of the functions. The result is one algebraic
equation per grid node, in which the variable value at that and a certain number of neighbor
nodes appear as unknown. Finite Difference method is applied to structured grids. These
grids consist of families of grid lines with the property that members of a single family do not
cross each other and cross each member of other families only once. The position of any grid
point within the domain is uniquely identified by a set of 2D indices.

Taylor series expansion or polynomial fitting is used to obtain approximations to the first and
second derivatives of the variables with respect to the coordinates. When necessary, these
methods are also used to obtain variable values at locations other than grid nodes
(interpolation).
FD method is especially easy to obtain higher order schemes on regular grids. The
disadvantage of FD method is that the conservation is not enforced unless special care is
taken. Cartesian grid for FD method is shown in Figure 2.1.
The idea behind finite difference approximations is borrowed directly from the definition of a
derivative:

Figure 2.1 : Cartesian grid for FD method

Using Taylor Series Expansions, approximate expressions are obtained for the first and
higher derivatives at point in terms of the function values at neighboring points as follows:

2.1

Another expression may be obtained by using Taylor Series Expansions at both and
7
 2.2

Here, H means “higher order terms”. Higher-order approximations of the first derivative can
be obtained by using more points to eliminate more of the truncation error terms. Using
is to obtain an expression for the second derivative at and substituting this expression in
equation 2.3, the following second order approximation is obtained:

2.3

2.3.5 Numerical grid

The discrete locations at which the variables are to be calculated are defined by numerical
grid which is essentially a discrete representation of the geometric domain. It divides the
solution domain into a finite number of sub domains (grid, elements, control volumes etc.)

• Continuous function is replaced by discrete function as grid or elements in a process

which is called discretization. Finite difference grid is shown below (Figure 2.2).

• Continuous time is approximated by discrete time steps.

Figure 2.2: Finite difference grid

2.3.6 Properties of numerical solution methods

The solution method should have certain properties. In most cases it is not possible to analyze
the solution method. One analyzes the components of the method; if the components do not
possess the desired properties, neither will the complete method but the reverse is not
necessarily true. Some important properties are given below

• Consistency

8
 • Stability

• Convergence

• Conservation

• Boundedness

• Realizibility

• Accuracy

(Truncation error)

Partial differential equation Finite differential equation

(consistency)

Boundary Condition

Exact solution (accuracy) Approximated solution

(convergencey)

Relation in analytical solutions

Relations in finite solutions

Fig. 2.3: Conceptual relationship between consistency, stability and convergence.

 Consistency

A finite difference scheme is said to be consistent or compatible when the truncation error
tends to vanish and the solutions to the finite difference equations approach the solution to
the differential equations as the computational mesh is refind ( i. e ∆x 0 and ∆t 0 ).
It is incompatible.

 Stability

9
The stability of a difference method concerns the amplification errors. These errors may be
introduced by inaccurate initial and boundary conditions. Stability is tested by asking if a
particular part of the solution is likely to amplify without limit until it destroys the
calculation.

The explicit and implicit finite difference schemes are subject to different stability criteria.
The following stability criterion can be discerned.

I. Courant-Friedrichs-Lawy criteria
II. Von-Neumann criterion

Courant-Friedriches-Lewy criterion

The condition for stability of an explicit finite difference scheme may be defined as

∆t ≤ ∆x/ C

Or | σ| = |c| ∆t/ ∆x ≤ 1

Where c is the celerity and σ is the Courtant number. This is called the Courtant Friedrichs-
Lewy (CFL) condition or simply the Courtant condition and was first established by the three
mathematicians in 1928. This is the usual criterion for the stability of explicit finite difference
methods.

Von-Neumann criterion

This criterion applies to both explicit and implicit finite difference schemes and states that the
amplitude of a particular wave should remain the same or decrease in time, i.e. should not
increase with time. This condition is expressed by a complex amplification factor ρ which
gives the relationship between the value of a variable at the time level n and that at the
advance time level (n+1), e.g.

Φ (j, n+1) = ϕ ρ (j, n)

The amplitude of the wave is multiplied by |ρ| and there is a phase shift equal to arg | ρ|and
there is a phase shift equal to arg (ρ)

where

| ρ|= [ ( real part of ρ)2 + (imaginary part of ρ)2]1/2

and arg (ρ) = tan-1 [ imaginary part of ρ/ real part of ρ]

Now if the absolute value of the amplification factor is less than unity for any value of the
wave length, the amplitude of the numerical solution will decreased and the scheme is
numerically stable. But when the amplification factor is greater than unity in absolute value.
The amplitude of the numerical solution will be multiplied by a factor greater than one at
each time step, causing any perturbation to grow without limit. In such a case the finite
10
difference scheme will be numerically unstable. Hence a necessary and sufficient condition
for stability is that

-1 ≤ ρ ≤ 1

| ρ| ≤ 1

Convergency

If the discrete solutions to the governing differential equations approach the exact solutions to
the governing equations approach the exact solutions as the computational mesh is refined,
i.e. as ∆x 0 and ∆t 0 , for a consistent finite difference scheme, then it is said to be a
convergent scheme, then it is said to be a convergent scheme. According to the Lax'
equivalence theorem, the necessary and sufficient conditions for a finite difference scheme to
be convergent are that the scheme is consistent and stable, i. e.

Consistency + stable = convergency

Accuracy

A finite difference method is said to be accurate if the solution of the finite difference
equation we obtain approaches the exact solution of the partial differential equation.

These are three ways of investigating the accuracy of the numerical methods:
1. By determining the truncation error
2. By analyzing the wave propagation and comparing the analytical and the
numerical solutions
3. By doing experiments with difference numerical parameters ( ∆x, ∆t etc).

2.3.7 Courant number

The quantity C= u∆t / ∆x is called Courant number. It is the ratio of time step ∆t to the
characteristics convection time, u/∆x, the time required for a disturbance to be convected a
distance ∆x. When courant number is less than 1.

2.4 iRIC SOFTWARE

The International River Interface Cooperative is an informal organization made up of

academic faculty and government scientists with the goal of developing, distributing, and
providing education for a public-domain software interface for river modeling. Formed in late
2007, the group released the first version of this interface, iRIC, in 2009. The purposes of the
activities of this project are creation of opportunities for interaction and provision and
exchange of information on issues to utilize knowledge and engineering about rivers to
support creation of more beneficial and sustainable river environments among administrative
engineers, construction consultant companies, river researchers and students focusing on
11
development of software for analysis of stream flows, river bed fluctuations and floods. The
iRIC software interface includes models for two- and three-dimensional flow, sediment
transport, bed evolution, groundwater-surface-water interaction, topographic data processing,
and habitat assessment, as well as comprehensive data and model output visualization,
mapping, and editing tools. All of the tools within iRIC are specifically designed for use in
river reaches and utilize common river data sets. The models are embedded within a single
graphical user interface so that many different models can be made available to users without
requiring them to learn new preand post-processing tools. iRIC provides a comprehensive,
unified environment in which data that are necessary for river analysis solvers (hereafter:
solvers) can be compiled, rivers can be simulated and analytical results can be visualized. The
highly flexible iRIC interface allows various solvers to be imported, or you can use one of the
iRIC solvers. Upon selecting the solver, iRIC selects functions suitable for the solver and
prepares the optimal simulation environment. Because the iRIC functions vary depending on
the solver, the method of using the iRIC application depends on the solver.

The iRIC software consists of three functions: preprocessor, postprocessor, and solver
(Figure 2.3).

Figure 2.4 : An outline of the iRIC Software, its functions and features.

2.4.1 Features of the flow field calculation model

 As a coordinate system, the general curvilinear coordinate system is adopted,
allowing direct consideration of complex boundaries and riverbed shapes.

 Calculations involving the confluence of a main channel and a tributary can be

performed.
12
  For the finite-difference method applied to the advection terms in equations of
motion, the user can chose the upwind difference method (primary precision) or the
CIP method (This is a high-order finite-difference method. By using a cubic
polynomial as an interpolation function, numeric diffusion is reduced, thus enabling
high-precision local interpolation).

 For the turbulent field calculation method, the user can select from among [Constant
eddy viscosity], [Zero-equation model] and [K-ε model].

 Various settings are possible for boundary conditions of the upstream and
downstream ends, including periodic boundary conditions, downstream end water
surface elevation setting and upstream end velocity setting. This makes it easy to set
boundary conditions from limited observation data.

 For setting the initial water surface profile, the user can select from among [Constant
slope], [Line], [Uniform flow calculation] and [Non-uniform flow calculation].

 The bottom friction evaluation method is set by using Manning's roughness

parameter. This parameter can be set to each computational cell.

 Any obstacle within the calculation target area can be taken into account on a
calculation-cell basis. For each calculation cell, a flag can be set to define an obstacle.
By this means, river structures such as bridge piers can be easily incorporated in
calculation.

 The effect of vegetation for the flow calculation can be introduced as a drag force.
The user can set the density of vegetation in each computational.

2.4.2 Features of the sediment transport and riverbed deformation calculation model

 The model allows the user to select between performing flow regime calculation only
and performing riverbed deformation calculation along with flow regime calculation.

 For sediment transport, the user can select between bed load only and bed load +
suspended load.

 For grain size distribution, the user can select between uniform and non-uniform.
When non-uniform grain size is chosen, changes in the grain size distribution in the
depth direction during calculation can be stored on a multilayer basis.

 The model can incorporate a slope collapse model. The unrealistic steep slope
sometimes can occur in the bed evolution computations, if the bed evolution is
simulated by using only Exner equation which expresses the mass conservation of
riverbed. In this model, if the angle of riverbed exceeds the critical angle, which is a
user parameter, the bed is instantaneously corrected as satisfying the critical angle. By
13
 introducing this model, you can treat the morphologic evolution processes such as
bank erosion, which cannot be well captured by using only continuity equation of
riverbed.

 Bank erosion takes into account of the tangent of angle of repose. More specifically,
when the riverbed slope exceeds the angle of repose as riverbed deformation
proceeds, adjustment is made by exchanging sediment with surrounding cells such
that the angle of repose is not exceeded. In addition, if the channel widens from bank
erosion, calculation grids are automatically moved accordingly.

 Hot Start is provided (continues calculation from where the previous calculation left
off, based on the previous calculation results).

2.5 BASIC EQUATIONS

2.5.1 Basic flow equations

2.5.1.1 Basic equations in orthogonal coordinates
The basic equations in an orthogonal coordinate system (x, y) before transformation
(mapping) into a general curvilinear coordinate system are as follows:

[Equation of continuity]

0 2.4

[Equations of motion]

2.5

2.6

in which
2.7

2.8

2.9

2.10

14
where h is water depth, t is time, u is velocity in the x direction, v is velocity in the y
direction, g is gravitational acceleration, H is water depth, τx is riverbed shearing force in the
x direction, τy is riverbed shearing force in the y direction, Fx is resistance by vegetation in the
x direction, Fy is resistance by vegetation in the y direction, Cf is riverbed shear coefficient, νt
is eddy viscosity coefficient, CD is drag coefficient of vegetation and as is area of interception
by vegetation per unit volume.

2.5.1.2 Transformation into general curvilinear coordinates

Next, basic equations of two-dimensional plane flow at orthogonal coordinates are
transformed into general coordinates (ξ, η). By transforming the equations into general
coordinates, it becomes possible to set a calculation mesh of any shape (in keeping with the
boundary conditions). The following describes how the equations can be transformed from
orthogonal coordinates into general curvilinear coordinates:

2.11

2.12

or,

2.13

where,
2.14

likewise,
2.15

2.16

or,

2.17

where,
2.18

Hence,

2.19

15
Here, assuming, we obtain,
2.20

Therefore,
2.21

Or,
2.22

2.23

Thus we obtain,
2.24

Letting the (ξ , η) components of velocity be (uξ , uη), then

2.25

2.26

or,
2.27

2.28

2.5.1.3 Basic equations in a general curvilinear coordinate system

Basic equations as obtained by transforming basic equations in an orthogonal coordinates
system (x , y) into a general curvilinear coordinate system are shown below:
[Equation of continuity]

2.29

[Equations of motion]

2.30

16
 2.31

where,
2.32

2.33

2.34

2.35

2.36

2.37

As for diffusion terms Dξ and Dη in the motion equation in general coordinates, since
developing those terms will make the number of terms huge, they are simplified by assuming
the following conditions:

The second-order derivative for the metric coefficient is assumed to be locally zero.
Those terms are locally treated as pseudo-orthogonal coordinates.
As a result, the diffusion terms can be approximated as follows:

2.38

2.39

where ξr and ηr are parameters each representing the ratio of the local grid size in general
coordinates to the full-scale length of the grid. They are defined as follows:

2.40

17
Note that to derive the approximate equations of Dξ and Dη above; the following relations are
used, based on the assumption of a relationship of local orthogonality.

2.41

2.42

2.43

2.44

where θ represents the angle formed by the x axis and the ξ axis (or the y axis and the η axis).

2.5.2 Turbulent field calculation method

Turbulence means disordered flow that contains eddies of various sizes and structures.
Nays2D allows the user to select the turbulent field calculation method from among
[Constant eddy viscosity], [Zero-equation model] and [k-ε model].

2.5.2.1 Constant eddy viscosity

Eddy viscosity coefficient vt is the apparent kinetic viscosity coefficient of a flow in
turbulent state. When [Constant eddy viscosity] is selected, calculation is performed with vt in
equations 3.5 and 3.6 being taken as 0.000001 m2/s.

2.5.2.2 Zero-equation model

Eddy viscosity coefficient vt is generally represented as the product of the turbulence
representative velocity vt and the representative length l.
2.45

For a flow field where depth and roughness change only slowly in the transverse direction,
assuming that the eddy viscosity coefficient in the horizontal direction and the eddy viscosity
coefficient in the vertical direction are on the same order, and considering that the bottom
friction velocity and the depth dominantly dictate momentum transport, the eddy viscosity
coefficient vt is expressed by the following equation.
2.46

where, α is a proportional constant.

Since α value related to momentum transport in the vertical direction, according to
experiments by Fisher and Webel and Schatzmann, is around 0.07, the eddy viscosity
coefficient νt is expressed using the Karman coefficient k (0.4) as per the following formula:
2.47

Since this modeling does not require any transportation equation for turbulence statistics, it is
called the "zero-equation model".

18
2.5.2.3 K-ε model
The eddy viscosity coefficient νt in the standard k-ε model is expressed by the following
equation:
2.48

where Cμ is a model constant. k and ϵ are obtained by the following equations:

2.49

2.50

Table 2.1 Model constants

Cμ C1ε C2ε σk σε
0.09 1.44 1.92 1.0 1.3

where C1ε, C2ε, σk and σε are model constants whose respective values are shown in Table 2.1
Note that Pkv and P εv are calculated with the following equations:

2.5.3 Bottom friction calculation method

In Nays2DH, bottom friction is set using Manning's roughness parameter. For Manning's
roughness parameter, the user can define this parameter in each computational cell.

2.5.3.1 Constant value

In Equation 2.4, riverbed shearing forces τx and τy are expressed by using coefficient of
riverbed shearing force Cf. The coefficient of riverbed shearing force Cf is estimated by
Manning's roughness parameter nm as follows:
2.51

This Manning’s roughness parameter can be estimated from the relative roughness height, ks,
by using the Manning – Strickler equation as follows.

2.52

19
where ks is the relative roughness height which is defines as d, d is a sediment diameter and g
is the gravitational acceleration.

2.5.4 Method of calculating resistance by vegetation

In Nays2D, resistance exerted by vegetation is set with the drag coefficient of vegetation CD
and the area of interception by vegetation per unit volume αs. The area of interception by
vegetation per unit volume αs can be set in each computational cell.
The area of interception by vegetation per unit volume αs is calculated with the following
equation proposed by Shimizu et al:
2.53

here ns is the number of vegetation, Ds is the average diameter of trunks and Ss is the
sampling grid width.

20
 Chapter 3
SIMULATION TECHNIQUE

3.1 INTRODUCTION

In this chapter, the simulation techniques for different test cases, details of flow domain and
flow parameters are presented. A brief description of 2D numerical model iRIC Nays2DH is
given and its operational procedure is explained.

3.2 SOFTWARE USED

In this study, iRIC Nays2DH is used to simulate the flow fields in an open channel with
groins. iRIC (International River Interface Cooperative) is a pre- and post-processing
software application and framework for computational models of flow and sediment-transport
in rivers. The application is a Graphical User Interface (GUI) that allows the model user to
build, run, and visualize the results of the system’s computational models. The GUI provides
tools for building both structured and unstructured grids, defining topography and other
boundary conditions on the grid, and defining grid-dependent values such as grain size,
vegetation, and obstacles by mapping measured values to the grid or by creating user-defined
polygons with attributes of grid dependent value. It combines the functionality of
MD_SWMS, developed by the USGS (U.S. Geological Survey) and RIC-Nays, developed by
the Foundation of Hokkaido River Disaster Prevention Research Center.

Nays2D is an analytical solver for calculation of unsteady 2D plane flow and riverbed
deformation using boundary-fitted coordinates within general curvilinear coordinates. The
solver's prototype was initially developed by Professor Yasuyuki Shimizu of Hokkaido
University in the 1990s. After many improvements, it was first adopted in 2004 as a
calculation solver for incorporation in the RIC-Nays riverbed deformation calculation pre-
post software of the Foundation of Hokkaido River Disaster Prevention Research Center
(Version 1.0).

Later refined by the inclusion of dynamic memory allocation by Ichiro Kimura of Hokkaido
University and outfitted with a bank erosion model by Yasuyuki Shimizu and a HotStart
function by Toshiki Iwasaki of Hokkaido University, Nays2DH (Version 2.0) was distributed
as one of the solvers included with iRIC in 2010 at the release of iRIC Version 1.0.

Further functional additions included the incorporation of a mixed-diameter multilayer model

proposed by Toshiki Iwasaki, and a river confluence model proposed by Takuya Inoue and
Michihiro Hamaki of Kaihatsu Koei Co. Ltd. Nays2DH was registered as a calculation solver
for iRIC Version 2.0 in March 2011 under the planning and supervision of Kazutake Asahi
from the Foundation of Hokkaido River Disaster Prevention Research Center (Version 3.0).
21
This model has an established reputation for calculation of unsteady flows accompanied with
turbulence and laminar flow and it is capable of dynamically showing the realistic motions of
unsteady eddies. In addition, its riverbed deformation calculations can reproduce the
generation, development and migration of sandbars with high precision. There have been
many examples of Nays2D being used on actual rivers for purposes such as assessing the
effects of trees and vegetation, making flood calculations, studying the effects of inflowing
rivers and simulating bank erosion disasters.
Nays is a model developed at Hokkaido University; the details of the model are described in
Shimizu (2002). The title of the model is not an acronym; it is an Ainu word meaning “small
river”. There are four primary features of Nays:

 It is two-dimensional ( vertically integrated).

 It is fully unsteady.
 It is cast in a general, nonothogonal coordinate system with variable cell size.
 It includes a much more sophisticated method for treating turbulence that includes both a
horizontal large eddy simulation and a suite of turbulence closures.

The non-orthogonal coordinate system allows more precise fitting of the coordinate system to
suit arbitrary channel curvature and variable width. More importantly, the more detailed
treatment of turbulence and large eddies allows predictions of time-variable behavior even for
steady discharges. Nays is currently the most sophisticated model within iRIC in terms of
handling advection of momentum and strong local unsteadiness. Nays also includes full
sediment-transport capabilities and morphologic change prediction, so the impacts of
unsteadiness on bed change can be assessed with this approach. Furthermore, Nays provides
a variety of particle-tracking information.

22
3.4 OPERATIONAL PROCEDURE OF SOFTWARE

The following are the basic procedures for operating Nays2DH with iRIC

Launching Nays2DH
Prepare to use Nays2DH with iRIC

Creating Calculation grids

Create a grid for calculation using river survey data and/or Digital Elevation Model
(DEM) data

Setting calculation conditions

Set simulation discharge, boundary conditions, roughness and other items

Making simulations
Use Nays2DH to run the simulation

Visualizing the calculation results

Visualize the simulation results, such as flow velocity, water depth and riverbed
Elevation, by using a contour map and/or vector map to see whether the simulation
Has successfully run

3.5 FLOW PARAMETER

The numerical tests consist four series, where each series contains several set of simulation
cases.
Series 1 : Simulation of 180º sharp and mildly curved bend.
Series 2 : Simulation of open channel flow with different orientation of groin.
Series 3: Simulation of flow and local scour around a single barb (groin) with different
orientations.
Series 4 : Simulation of flow and local scour around a group of three groins.
The hydraulic parameters for 180º bend channel are shown in Table 3.1. The parameters for
case of 180 degree sharp bend were taken same as the experiments by Rozovskii (1961). The
parameters for case of 180 degree mild bend were taken same as the experiments by De
Vriend (1979).

23
Table 3.1 : Hydraulic parameters for numerical simulation in Series 1 are given bellow
( Case S1C1 is the case for sharp bend and S1C1 is for mild bend)
L (m) W, Qo, ho, m Bottom Δt, R/B n Courant
Case U/S D/S m m3/s slope, sec number
no. So

S1C1 6 3 0.8 0.0123 0.0600 0 0.005 1.0 0.0102 0.04

S2C1 6 6 1.7 0.0671 0.1953 0.016 0.020 3.5 0.0350 0.0149

∆t = Time step, L = Length of the straight portion, Q0 = Upstream discharge, h0 = Mean

depth, So = Bottom slope, n = Manning’s coefficient.

Table 3.2 Hydraulic parameters for numerical study in series 2 are given bellow
(S2C1 is for 45º groin, S2C2 is for 90º groin and S2C3 is for 135º groin)

Groin L(m)
Case Position U/S D/S W (m) Q 0 (m3/s) h0(m) S0 ∆t (sec)
no

S2C1 45º 2 4 0.9 0.0430 0.189 0.001 0.003 0.01

S2C2 90º 2 4 0.9 0.0430 0.189 0.001 0.003 0.01
S2C3 135º 2 4 0.9 0.0430 0.189 0.001 0.003 0.01

Here, L = Length of the straight channel, W=Width of channel, Q0 =Upstream discharge, h0 =

Water depth at downstream, n = manning’s roughness co-efficient.
The parameters for case-S2C1 for the flows in an open channel with groin of 450 angled to
the direction of flow were taken same as the experiments by Rajaratnam and Nwachukwu
(1983) and in Table 3.2. The hydraulic parameters for the 900 and 1350 angled groin are the
same as the case-S2C1.

The hydraulic parameters for the bed evaluation in an open channel with 45º angled barb
were taken as same as the numerical model presented by Hossain et al ( 2012) and shown in
Table 3.2. Hydraulic parameter for both 90º and 135º angled barb were taken same as 45º
angled barb.

24
Table 3.3 : Hydraulic parameters for numerical study in series 3 are given bellow
(Case S3C1 is for without barb, S3C2 is for 45º stream barb, S3C3 is for 90º stream barb, S3C4
is for 135º stream barb and S3C5 is for barb at both sides)

Case Groin L (m) B(m) S B/D D(m) Q U d50

no position (m/m) (m3/s) m/s mm
S3C1 No 13.3 0.8 0.005 18 0.044 0.028 0.64 0.77
barb
S3C2 45º 13.3 0.8 0.005 18 0.044 0.028 0.64 0.77
S3C3 90º 13.3 0.8 0.005 18 0.044 0.028 0.64 0.77
S3C4 135º 13.3 0.8 0.005 18 0.044 0.028 0.64 0.77
S3C5 Double 13.3 0.8 0.005 18 0.044 0.028 0.64 0.77
barb

Here, L = length of the channel, B = width of the channel, S = slope of the channel, D = water
flow depth at downstream, Q = Discharge of the channel, U = velocity, Fr = Froude number.
d50 = sediment size.

The flow parameters for three parallel spur dikes in a straight open channel were taken same
as both the experiment and numerical study presented by karami et al (2014) in table 3.4.

Table 3.4 Hydraulic parameter for numerical study in series 4 are given bellow
(Case S4C1 is for three spur dikes with discharge 0,035 m3/s and S4C2 is for three spur dikes
with discharge 0.046 m3/s).

Case no. L B Y Q U Fr d50

m m m m3/s m/s - mm
S4C1 5 1 0.15 0.035 0.233 0.19 0.91
S4C2 5 1 0.15 0.046 0.307 0.25 0.91

Here, L = length of the channel, B = width of the channel, S = slope of the channel, Y = water
depth, U = velocity of the channel, Fr = Froude number, d50 = Sediment size.

3.6 COMPUTATIONAL SCHEMES AND FLOW DOMAIN

The governing equations for mean velocities and turbulent flows are discretized with the
finite difference method based on full staggered boundary fitted coordinate system. The basic
equations are discretized as fully explicit forms and solved successively with the time
increment in step by step. It is solved using iterative procedure at each time step. Constant
discharge at upstream and constant depth with zero velocity gradients was given as
downstream boundary conditions. Finite volume method was used for the solution of the
equations. 2D grid, Q, he and initial flow conditions were given as input.

25
 3.6.1 Series 1

Figure 3.1 shows the sketch of the flow domain for 180o sharp bend which was performed
under the same conditions of the experiments conducted by Rozovskii (1961). Figure 3.2
shows the sketch of the flow domain for 180o mild bend which was performed under the same
conditions of the experiments conducted by De Vriend (1979). For the first Case 163 x 25
and for the second Case 126 x 25 grids were used with moveable bed condition.
Transverse distance, y ((m)

Longitudinal distance, x (m)

Figure 3.1: Sketch and grid of an open channel with 1800 sharply curved bend (Case S1C1).

26
 Transverse distance, y (m)

Longitudinal distance, x (m)

Figure 3.2: Sketch and grid of an open channel with 180o mildly curved bend prepared by
iRIC software (Case S1C2).

3.6.2 Series 2

Figure 3.3 shows the sketch of the flow domain for flows in an open channel with groin of
900 angled to the direction of flow which was performed under the same conditions of the
experiments conducted by Rajaratnam and Nwachukwu (1983). Figure 3.4 and Figure 3.5
show the sketch of the flow domain for 45o and 135o which were performed under the same
conditions as case-1. For all these cases 120 x 20 grids were used and the bed of the channel
was fixed.

27
 y

0.9m

2m 4m x

Figure 3.3 : Sketch and groin orientation for 45º angled to the flow direction (Case S2C1).

0.9m

2m 4m x

Figure 3.4: Sketch and groin orientation for 90º angled to the flow direction ( Case S2C2).

28
 y

0.9m

2m 6m x

Figure 3.5: Sketch and groin orientation for 135º angled to the flow direction (Case S2C3).

3.6.3 Series 3

Figure 3.7 shows the sketch of the flow domain for flows in a straight open channel with
groin of 450 angle to the direction of flow which was performed under the same condition
conducted by Hossain et al (2012).

Figure 3.8 and Figure 3.9 show the sketch of the flow domain for 90o and 135o which were
performed under the same conditions . For all these cases 100 x 10 grids were used and the
bed of the channel was moveable.

0.8m

13.3m

Figure 3.6: Sketch and Preparation of grid of a straight channel by iRIC Software (Case
S3C1)

29
 y

0.8m Flow

13.3m

Figure 3.7: Sketch and barb installation for 45º angled to the flow direction prepared by iRIC
software (Case S3C2).

0.8m Flow

13.3m

Figure 3.8: Sketch and barb installation for 90º angled to the flow direction prepared by iRIC
software (Case S3C3).

30
 y

Flow
0.8m

13.3m

Figure 3.9 : Sketch and barb installation for 135º angled to the flow direction prepared by
iRIC software (Case S3C4).

Figure 3.10: Sketch and barb installation for double barb on both side prepared by iRIC
software (Case S3C4).

3.6.4 Series 4

Figure 3.11 shows the sketch of the flow domain for a straight channel with three
perpendicular groin which was based on experimental study presented by karami et al (2014).
In this case, 120 x 26 grid were used.

31
 5m

3.16 1m

Figure 3.11 Sketch and grid preparation of a straight channel with three spur dikes

32
 Chapter 4
. SIMULATION OF FLOW AT A BEND

4.1 INTRODUCTION
2D numerical model, iRIC Nays2DH, is used to simulate a 180º sharply and 180º mildly
curve bend to assess the predictability of the model for the flow field having strong secondary
current. The numerical simulation has been performed on fixed bed condition and the
simulated velocity field and water surface profile are compared with previous experimental
result to evaluate the capability of the model.
Natural river are characterized by complex geometrics such as channel expansions and
contractions, bends and variable bed and bank variable bed and bank roughness. In the field
of river engineering, it is important to understand the flow behavior exits at sharp bends.
Because of having river's highly turbulent in nature, it is difficult, time consuming and costly
in order to obtain all necessary details by lab experiment and field observation. So, numerical
simulation can be better option to get all necessary details within a short time. The
computational fluid dynamic (CFD) is branch of fluid dynamics which provides fast, accurate
and flexible fluid flow simulation. Computational fluid dynamics is based on Navier-Stokes
equation. 3D hydrodynamic model is required in order to accurately simulate the flow in a
meandering channels. For this purpose several 3D models have been developed by Wang and
Hu (1990), de Vriend (1980) and Shimizu et al (1990) to simulate the flow field in
meandering channels. To avoid some practical engineering problems, such as alluvial
geomorphic engineering problems, such as alluvial geomorphic processes, to use 3D model is
not efficient. Flow passing through meandering channels also has influence of the secondary
flow in which there is a transverse circulation introduced by the centrifugal force.
Measurement by Rozovskii (1961) and de Vriend (1979.1980, 1981) have shown that the
secondary flow near the water surface moves toward the outer bank and that near the bed
moves toward the inner bank.

As shown in Figure 4.1, the two-dimensional flow has only two velocity components; they
are longitudinal or stream wise and transverse or width wise velocities. Vertical or depth wise
velocity component is ignored. At the straight portion, the width-wise variation of flow is
very small, and only the stream wise velocity or the longitudinal velocity dominates the flow.

33
Figure 4.1: Definition sketch of the flow pattern at the straight portion of the river.

At the bend, the flows of water are dominated by both the longitudinal and transverse
velocity components as shown in Figure 4.2. Secondary current is generated due to the
centrifugal force at the bend when water enters into the bend. Though the water depth is
equal for the inner bank and outer bank at the straight portion, it differs from inner bank to
outer bank at the bend of the river. This difference in water level results from the transverse
slope caused by the centrifugal force. It causes a higher velocity at the inner bank than the
outer.

Figure 4.2 : Definition sketch of the flow pattern at the bend of the river.

To simulate meandering channel flow by 2D models, generally the depth-averaged Navier–

Stokes equations numerically by using finite element technique and the simplified depth-
averaged Navier–Stokes are used. The momentum transfer due to the secondary flow is
inadequately accounted for by the depth averaged models and various empirical functions
34
have to be introduced in these models to approximate the secondary flow. In many practical
cases, the effect of secondary flow is not significant when the channels are not curved or the
curvature effect is small. This study employed the Cartesian coordinate so that the
hydrodynamic model can be easily applied to both meandering and non-meandering
channels.

4.2 SIMULATED RESULTS OF 180o SHARPLY CURVED CHANNEL (Case S1C1)

The flow domain consists of 0.8 m wide channel having a rectangular cross section. The
straight reaches at the upstream has an effective length of about 6.0 m and downstream has an
effective length of about 3.0 m. The internal and external radius of curvature is 0.4 m and 1.2
m respectively. The ratio of radius of curvature to channel width is 1.0. The cross section of
the channel bend is a 0.8 m wide rectangle, and entire channel bed is horizontal. The channel
bottom is made smooth. The discharge at the inlet is 0.0123 m3/s, the averaged mean velocity
is 0.265 m/ s, and the averaged flow depth is 0.06 m. The computational mesh is 163 x 25
with 163 cross sections in the longitudinal direction and 25 nodes in each cross section. The
logarithmic law at the sidewalls was used in the numerical simulation. The simulation reach
consists of the U bend and the effective straight reaches at the upstream and downstream.
Constant discharge at upstream and constant depth with zero velocity gradients was given as
downstream boundary conditions. Sharply curved channel is simulated of Standard Solver
type with finite differential method and discretizing the advection terms as upwind scheme by
iRIC Nays. iRIC Nays results are compared with experimental results.

Figure 4.3 (a) shows the measured contours of water surface of Rozovskii’s (1961)
experiment. Figure 4.3 (b) shows the simulated result. It can be seen that the numerical model
successively reproduces the higher water surface elevation along the outer bank than the
inner bank. So the overall agreement between measured and simulated profiles appears to be
quite good.

The longitudinal distribution of water surface along the inner bank and outer bank are
compared in Figure 4.4. The horizontal axis in the figure is the dimensionless distance from
the entry of the channel bend. From this Figure it is seen that, the water surface elevation at
the outer bank is much higher than that in the inner bank throughout the bend. It can be also
seen that the computed flow depth is close to the measured data and the agreement between
measured and simulated profiles is satisfactory.

35
 (a) (b)

Figure 4.3: Distribution of water surface elevation for Case S1C1 (a) Experiment by
Rozovskii (1961), (b) Simulated by iRIC Nays.

Figure 4.4 : Comparison of simulated water surface profile with experimental results for 180o
sharply curved channel.

36
 (a) (b)
Figure 4.5: Comparison of width wise variation of velocity vector at different section of the
180o sharply curved channel (Case S1C1) (a) Experimental by Rozovskii (1961), (b)
Computation by iRIC Nays2DH.

Figure 4.5 (a) present the measured result by Rozovskii (1961) and 4.5 (b) shows the field
predicted velocity vector field in the channel bend. The simulated result by iRIC Nays is
found to be well compared with the experiment. For both the cases the velocity of flow is
greater in the inner section than the outer section. The velocity in the inner bend is found to
be decreased when the flow enters the straight portion after the bend. From experimental and
iRIC Nays result, it is observed that in the straight portion at outlet, the outer bank velocity is
higher than inner bank. For a sharp bend, the radius of curvature is in the order of the width
of the channel, thus the effect of secondary flow is very strong.

4.3 SIMULATED RESULT OF 180º MILDLY CURVED CHANNEL ( Case S1C2)

In this section, the experiment by De Vriend (1979) is numerically reproduced. The channel
consists of 1.7 m wide flume having a rectangular cross section. The bottom of the channel is
horizontal and the sidewalls are vertical. The radius of curvature of the flume axis in the

37
bend is 4.25 m and the upstream and downstream straight reaches have an effective
length of about 6.0 m. The ratio of radius of curvature to the channel width is 3.5. The
discharge at the inlet is 0.0671 m3/s, the averaged velocity is 0.202 m/s and the averaged flow
depth is 0.1953 m. The logarithmic laws at the sidewalls were used in the numerical
simulation. The simulation reach consists of the U bend and the effective straight reaches at
the upstream and downstream. Constant discharge at upstream and constant depth with zero
velocity gradients was given as downstream boundary conditions.

Figure 4.6 plots the simulated flow field as velocity vector field and the surface elevation in
shaded color. The velocity vectors at inlet and outlet straight portion are found to be uniform.
The velocity along the inner bank is higher than the outer bank at the bend. Water surface
elevation at inlet is higher than the outlet straight portion. At the bend, the water surface
elevation along the outer bank is higher than the inner bank.

Surface Elevation(m)

(a) (b)

Figure 4.6: Velocity vector super imposed on the surface elevation contour (Case S1C2) (a)
Experiment by De Vried (1979) (b) Simulation by iRIC Nays2DH.

The comparison of the simulated velocity with previous experiments is shown in Figure 4.7.
Figure 4.7 (a) shows the experimental velocity distribution and Figure 4.7 (b) shows the
computed velocity distribution. The comparison shows good agreement both at bend and
straight portion of the channel. It is observed that the velocity vector along the inner bank is
longer than the outer bank. The acceleration of the stream wise velocity at the inner bank is
due to the transverse convection of momentum induced by the secondary current. However,
for this mild bend, the radius of channel curvature is much larger than the width of the
channel and thus the effect of secondary flow not very strong.

38
 (a) (b)

Figure 4.7: Distribution of depth-averaged velocity 180o mildly curved channel (Case S1C2)
(a) Experiment by De Vriend (1979), (b) Simulation by iRIC Nays2DH.

Figure 4.8: Water surface profile of 180º mildly bend channel (S1C2) .

39
The comparison of simulated water surface elevation at the inner bank, centerline, and outer
bank was plotted in Figure 4.8. From the Figure 4.8, it is clear that the water surface elevation
at the outer bank is much higher and at the inner bank is much lower than that of center of the
channel water surface elevation throughout the bend. The rise of flow at the outer bank
results from the centrifugal force. This change in water surface elevation starts from just
before to the entrance of the bend and continued up to just after the exit of the bend.

40
 Chapter 5
SIMULATION OF FLOW FIELD AROUND GROIN

5.1 INTRODUCTION
2D numerical model Nays2DH is used to simulate the flow field in a straight open channel
with 45º, 90º and 135º groin angled with the flow direction. Depth-averaged k-ε model is
used as turbulence closure model with finite differential advection as upwind scheme during
simulation. The simulation has performed to evaluate and analysis the velocity profile and
bed shear stress profile.

5.2 NUMERICAL TEST

5.2.1 Flow domain and hydraulic parameters

The hydraulic parameters for different test cases that were simulated shown in Table 5. 1.
The parameters for case 1 of flow in a straight open channel with groin of 90° angle to the
direction of flow were taken the same as the experiments by Rajaratnam and Nwachuku. In
their experiment, a 37 m long, 0.9 m wide and 0.76 m deep flume with smooth bed and sides
were used. The groin was an aluminum plate with 3 mm thickness and 152 mm length and
was projected partly above the water surface. The hydraulic parameter for the 45° and 135°
angled groin are the same as case 1.

Figure 3.3 in Chapter 3 shows the sketch of the flow domain for flows in a straight open
channel with groin of 90° angled to the direction of flow which was performed under the
same conditions of the experiment conducted by Rajartnam and Nwachukwu. Figures 3.4 and
3.5 in chapter 3 show the sketch of the flow domain for 45° and 135° which were perfomed
under the same condition as case 1. For all these case, 120 * 20 grids were used considering
fixed bed condition. Figures 4 shows a sample of computational mesh that consists of 120
grids in longitudinal direction and 20 grids in transverse direction.

For all the cases, the computational domain was 6 m in length and 0.9 m in width. The
upstream and downstream boundaries were located at 2m and 4m away from the groin,
respectively. The constant discharge at upstream was 0.043 m3/s, and at the downstream the
constant flow depth was 0.189 m. Manning,s roughness coefficient n = 0.01 and the time step
∆t =0.003 sec are assigned in the present computation. The simulation are performed in fixed
bed condition with bottom slope, S0 = 0.001.

Quanhong and pengzhi also used the same experimental data of Rajaratnam and Nwachukwu
to validate their model. Governing equations of their model were shallow water equations and
depth-averaged k- model as the turbulence model. According to Quanhong and pengzhi
41
model, the computational domain was 6 m in length and 0.9 m in width. Flow flux of
0.047817 m2/s and water depth of 0.189 m are specified at the upstream and downstream
boundaries, respectively. The simulation is performed under movable bed condition.

Table 5.1 : Model constant for nonlinear k-ε model

Cases Groin L (m) W (m) Q0 ( m3/s) h0 (m) S n

Position U/s D/s

Case S2C1 45° 2 4 0.9 0.043 0.189 0.001 0.01

Case S2C2 90° 2 4 0.9 0.043 0.189 0.001 0.01
Case S2C3 135° 2 4 0.9 0.043 0.189 0.001 0.01

Here, L = length of the straight channel, w = width of channel, Q0 = upstream discharge, h0

= water depth at downstream, S0 = channel slope, and n = Manning's roughness coefficient.

(a)

(b)

(c)
Figure 5.1: Computed velocity vector for (a) case S2C1, (b) case S2C2, (c) case S2C3
developed by iRIC Nays2DH.

From the simulated results it is seen that the flow is deviated towards the left bank than the
right bank due to obstruction of flow by groin.

42
5.2.2 Computational Scheme

The governing equation for mean velocities and turbulent flows were discretized with the
finite difference method based on full staggered boundary fitted coordinate system. The basic
equation are discretized as fully explicit forms and solved successively with the time
increment step by step. It is solved using iterative procedure at each time step. Constant
discharge at upstream and constant depth with zero velocity gradients were given as
downstream boundary conditions. Finite difference scheme was used for the solution of the
equations, where the nonlinear convection terms were discretized with the upwind scheme.

5.3. SIMULATION FLOW FIELD

Figures 5.2 show the predicted velocity vector of the flow field around the groin for different
angles with the flow. For all the cases, the model is found to reproduce the general flow
features of the flow field around the groin successfully. From the simulated results, it is seen
that at the upstream boundary the flow is uniform and hence the flow vectors are straight and
parallel to bank. However, when the flow approaches the groin, the flow is deviated towards
the left bank compared to the right due to obstruction of flow by the groin. A recirculation
zone is observed just downstream region of the groin for all the cases. Figure 5.2 depict that
for case S2C1 (45° groin) the deviations of the vectors are higher than the case S2C2 (90°
groin). 135° groin for case S3C3 shows the least deviation compared to other cases.

(a)

(b)

43
 (c)
Figure 5.2: Computed recirculation around groin for (a) case S2C1, (b) case S2C2, (c) case
S2C3 developed by iRIC Nays2DH.

5.3.1 Streamline
Figure 5.3 shows the streamline around a groin, which is obtained by the present study. From
the streamline contour it is found that at the position of groin head the flow is deviated
towards the opposite bank and at the downstream of the groin there is a circulation. Figures
6.3 depict that case S2C1 ( 45° groin) the deviations of streamline are higher compared to
case S2C2 ( 90° groin). 135° groin for case S2C3 shows the least deviation compared to the
other cases. The length of circulation zone at downstream of the groin is also higher for case
S2C2 and lower in case S2C3.

(a)

(b)

( c)

Figure 5.3: Streamline around a groin for (a) case S2C1, (b) case S2C2 and (c) case S2C3
developed by iRIC Nays2DH.

44
5.4 COMPARISON OF SIMULATED RESULT WITH PREVIOUS STUDIES

Simulated results such as computed velocity profiles and bed shear stress profile for 90°
groin (case S2C2) are compared with previous experimental as well as available previous
simulation results. The comparison is described in the following sections.
5.4.1 Computed Velocity profiles

For 90° groins, the resultant velocity profiles are compared with available experimental result
measured at y/l = 1.0,1.5,2.0,3.0 and 4.0, where l = 152mm is the length of the groin. Figure
8 shows the comparison of velocity profiles for different lateral distance (y/l). Here, the initial
flow velocity profiles for different lateral distance (y/l). Here, the initial flow velocity and
water depth were U0 = 0.253 m/s and H = 0.189m, respectively. In the figure, all the
velocities were normalized by U0. In the figure, x/l = 0 indicated the groin position along
flume direction.
Good agreements are found between experimental and calculated results. For the velocity
profiles, the only large discrepancy occurs at y/l = 2. Where the calculated results
underpredict the experimental data in the downstream zone of the groin. It is noticed that the
numerical result reported by Quanhong and pengzhi and Sarveram et al. also underpredicted
the experimental data largely in this region. This may be due to the very high velocity
gradient arising in this region which makes the depth-averaged model inapplicable.
Otherwise, the possible reason may come from the experimental measurement errors in this
region.

(a)

45
(b)

(c )

46
 ( d)

( e)

Figure 5.4: Comparison of resultant velocity profiles (W) with the available previous studies
for 90° groin (Case S2C2).

5.4.2 Computed Bed shear stress profiles

Figure 5.5 shows comparison of bed shear stress profiles (τ) of predicted results with the
available previous studies. In the figure, all the shear stresses are normalized by τ0 = 0.1293
N/m2. The comparison between the numerical results and experimental data shows reasonable
agreement. For the lateral sections up to y/l = 2, the simulated results underpredict the
downstream shear stress slightly. The numerical results reported by Quanhong and pengzhi
also show similar prediction. It may be the deficiency of the depth-averaged model to account
for the high velocity gradient arising near to the groin.

47
 (a)

(b)

( c)

48
 ( d)

(e)

Figure 5.5: Comparison of bed shear stress profiles (τ) of predicted results with the available
previous studies (Case S2C2).

5.5 SIMULATION RESULT FOR DIFFERENT CASES

5.5.1 Predicted velocity profiles

Figure 5.6 shows the predicted velocity profiles (W) of present study for case S2C1, case
S2C2 and case S2C3. For each case, the longitudinal velocity profiles are compared at a lateral
position of y/l =1.0, 1.5, 2.0, 3.0 and 4.0. In each case, the longitudinal velocity profiles are
compared at a lateral position of y/l = 1.0, 1.5, 2.0, 3.0 and 4.0. In the figure, all the velocities
49
are normalized by U0 =0.253 m/s and here x/l = 0 indicates the groin position in longitudinal
direction. The comparisons among the three cases show reasonable differences. Along line y/l
=1, the peak of velocities is found at longitudinal distance of x/l = -0.28, 0.0 and 0.56 for case
S2C1, case S2C2 and case S2C3, respectively. For y/l = 1.0 in all the cases, in all the cases, the
velocity is found to be decreased from upstream velocity just before the position of groin
head. In other words, in addition to the circulation zone at downstream of groin, a small low
velocity zone is observed just upstream of groin at right bank. Along the line y/l = 1.5, the
peak of velocities are found to be decayed with the longitudinal distances of x/l = -1.0, 0.22
and 0.67 for case S2C1, case S2C2 and case S2C3, respectively. The peak of velocities are
found to be decayed with the longitudinal distance x/l . For y/l = 2, the peaks of the velocities
are found at a distance of x/l = -0.11, 1.22 and 2.54 for case S2C1, case S2C2 and case S2C3,
respectively. For y/l ≥ 2, the decay of velocity along the downstream is found to be slower
compared to the region y/l ≤ 1.5.Along the line y/l =3, the peak of the velocities is found at
longitudinal distance of x/l = 2.87, 3.53, and 3.88 for case S2C1, case S2C2 and case S2C3,
respectively. Along the line y/l = 4, the peak of velocities is found at longitudinal distance of
x/l = 3.33, 3.89, and 4.44 for case S2C1, case S2C2 and case S2C3, respectively.

From the figures, it is observed that the peak of the velocity is found maximum velocity is
found lower than the other two cases for all the section y/l. However, W/U 0 for 90° groin is
found maximum for y/l ≤ 2; and for y/l > 2 , the velocity is maximum for 45° angled groin.

(a)

50
(b)

(c)

51
 (d )

(e)

Figure 5.6: Comparison between resultant velocity profiles (W) of present studies for case
S2C1, Case S2C2 & Case S2C3 along lateral distance (a) y/l=1, (b) y/l =1.5, (c) y/l=2 ,(d)
y/l=3, (e) y/l=4.

5.5.2. Predicted bed shear stress profiles

Figure 5.7 shows the predicted bed shear profiles (τ) of the proposed model for Case S2C1,
case S2C2, and case S2C3. For each case, the longitudinal shear stress profiles are compared at

52
a lateral position of y/l = 1.0, 1.5, 2.0, 3.0, and 4.0. In the figure, all the shear stresses are
normalized by τ0 = 0.1293 N/m2, which is the value of upstream region. Here, x/l =0 is the
groin position along longitudinal direction. The profiles of bed shear stress are found in
similar pattern of velocity profiles. For different values of y/l, the peak of shear stresses is
found at the same longitudinal distance (x/l) as reported in velocity profiles. For y/l = 1.0 in
all the cases, the shear stress is found to be decreased from its upstream value just before the
position of groin head.

From the figures, it is observed that the peak of bed shear stress is found maximum at the
position of head of head of groin when y/l =1. However, the position of maximum bed shear
stress is found to be shifted towards downstream with increasing y/l. These peak shear stress
are found to be shifted downstream with increasing y/l. These peak shear stresses are found to
be decayed with the longitudinal distance x/l. However, for y/l ≥ 2, the decay of shear stress
along the downstream is found to be slower compared to the region y/l ≤ 1.5. The maximum
bed shear stress for case 3 is found lower than the other two cases for all the sections of y/l.
However, τ/ τ0 for 90 ° groin is found maximum for y/l ≤ 1.5, and for y/l ≥ 2, the velocity is
maximum for 45° angled groin.

(a)

53
 (b)

Case S2C1
2.5 Bed shear stress along y/l =2 Case S2C2
2 Case S2C3

1.5

0.5

0
-6 -4 -2 0 2 4 6 8 10
x/l
( c)

(d)

54
 (e)

Figure 5.7 : Comparison between computed bed shear stress ( ) of the proposed model for
case S2C1, case S1C2, and case S3C2 along lateral distance, y/l = 1, 1.5, 2, 3, and 4.

5.6 COMPUTED VELOCITY CONTOUR

Figure 5.8 shows the simulated velocity contour in shaded color for case S2C1, case S2C2 and
case S2C3, respectively. The results are found to be similar in nature. For all the cases, the
velocity contour along the left bank is higher than right bank at the downstream region of the
groin; it indicates the deflection of flow towards right bank and sheltering of flow by groin at
east bank. The downstream circulation zone is found to be larger for 45° groin compared to
90° groin; similarly the circulation zone for 90° groin is found to be larger compared to 135°
groin.

In the figure, the flow field is seen to be divided into four distinct regions: (I ) the uniform
flow at upstream end, (II) big circulation of low velocity zone at the downstream of the groin
created due to sheltering of groin where the velocity is zero in the center of the circulation
(III) a small low velocity zone near the foot at the upstream of groin, and (IV) high velocity
zone opposite to the circulation zone created due to the deflection flow. The high velocity
zone is found to be decayed towards downstream and the low velocity in circulating zone is
found to be regained.

(a)

55
 (b)

(c)

Figure 5.8: Velocity contoured around a groin for ( a) case S2C1 (b ) case S2C2, and (c ) case
S2C3 at the end of simulation by iRIC Nays2DH.

5.7 COMPUTED SHEAR STRESS

Figure 5.9 shows the simulated bed shear stress contour in shaded color for case S2C1, case
S2C2 and case S2C3, respectively. The results are found to be similar in nature. Like the
velocity contour, for all the cases, the shear stress contour along the left bank is higher than
the right bank at the downstream region of the groin. Four distinct flow regions as described
in velocity contour are also clearly visible in the contour of shear stress.

56
 ( a)

(b)

(c)
Figure 5.9 : Bed shear stress contour around a groin for test (a ) case S2C1 (b) case S2C2 and
c) case S2C3 at the end of simulation by iRIC Nays2DH.
.

57
 Chapter 6
LOCAL SCOUR AROUND GROIN

6.1 INTRODUCTION
iRIC Nays2DH numerical model is also applied to simulate an open channel with single barb
with different orientation to observe the effect of barb on flow. Modeling of flow near barb is
important, especially associated with developing scour hole and deposition of sediment. The
numerical simulation has performed to evaluate the streamline variation, velocity contour
line, bed change contour, bed elevation and depth contour diagram in the presence of barb
and without barb. Besides, single barb, double barb has also studied with adequate hydraulic
parameter. In this study, the numerical tests by Hossain et al (2012) is reproduced. He called
groin as barb. Therefore, In this chapter barb is used instead of groin.

6.2 BED DEFORMATION AROUND SINGLE GROIN

Hossain et al (2012) has done an numerical simulation on a straight channel to observe the
effect of barbs on channel bed configuration. Barbs are used to protect banks for meandering
river as well as straight river. The primary function of barbs is to deflect the strong main flow
away from critical zones towards the channel center and prevent erosion of banks (Hansen
and winter, 1996). It also used on long reach of relatively straight or gently curving banks
that need protection. The approaches to protect stream banks from erosion during periods of
high flows including various bio-technical type channel revetments (Schiechtl and Stern,
1997). The length of this straight channel is 13.3m and wide is 0.8m which is generated by
iRIC-Nays2D. The channel shape was considered as rectangular flat movable bed with
simple cross section. The slope of channel is 0.005, the model is designed by subcritical flow.
There are hundred grids in longitudinal direction and ten grids in lateral direction,
respectively. The change of width of the grid channel in the direction of flow is constant.
The model was run two settings: first, before constructing the barb and second, after
constructing the barbs. At first flow approached into the channel pre-existing condition and
find the velocity vector, flow depth and the elevation of the bed configuration, respectively.
In this simulation, barb is used as a obstacle at one channel sidewall and it is 4.5m apart from
the upstream. The barb length is taken 24 cm which is one third of the channel width. A
rectangular riprap shape barb is formed at 45° angle with channel bank.

According to NRCS (2005), each barb extends from the bank-line to the proposed thalweg
location. Generally the length of the barb not exceed one-third of the cross-section top width
at bankfull stage, as Matsuura and Townsend (2004) observed in their study that, the overall
length should not exceed about one quarter the channel-forming flow width. By decreasing
the nodes elevation point at the barb end, the structure set as a sloping ridge from the bank to
the main channel flow. The barb placed at an angle of 45° with the upstream bank of the
58
channel as NRCS (2005) state the horizontal angle between the tangential line placed along
the upstream bank and the centerline of the longitudinal axis of the barb have varied from 30°
to 60°.

The generated grid channel was used to set the calculation condition and simulation. The
symbol L represents as the total length, B is the internal width and S is the slope of the
channel, D is the normal flow depth, Q is the discharge, U is the velocity of the flow, B/D is
the width to average flow depth ratio, Cf is the roughness coefficient correspond to d50, where
d50 is the mean sediment diameter and Froude number, Fr =U /√gD, here g is gravity
acceleration.

Table 6.1: Detail flow calculation condition ( for all the cases of series)

Cases Barb L B S B/D D Q U Fr D50

angle m m m/m - m 3
m /s m/s - mm
S3C1 No 13.3 0.8 0.05 18 0.044 0.028 0.64 0.977 0.77
Barb
S3C2 450 13.3 0.8 0.05 18 0.044 0.028 0.64 0.977 0.77
0
S3C3 90 13.3 0.8 0.05 18 0.044 0.028 0.64 0.977 0.77
0
S3C4 135 13.3 0.8 0.05 18 0.044 0.028 0.64 0.977 0.77
S3C5 Two 13.3 0.8 0.05 18 0.044 0.028 0.64 0.977 0.77
Barb

In the calculation condition the professional solver type with bed deformation was used. The
finite differential methods of advection term was used as CIP method. For calculation of
numerical simulation the cyclic boundary condition was used with initial water surface as
uniform flow. The upstream velocity and water surface at downstream was also considered as
uniform flow. Constant discharge time series were used from the beginning up to 5 hours.
Output time interval was selected as 10sec. the calculation time was set according to Courant-
Friedrichs-Lewy (CFL) condition as stated in equation (3). The CFL condition is necessary
for convergence of flow while solving certain partial differential equations numerically by the
method of finite difference (Courant et al. 1956).

∆t ≤ C ∆t = C * ∆x / u (3)
∆x/u

Where C is a dimensionless constant and it's vary from 0.1 ~ 0.3. The maximum iteration
times of water surface calculation was set as 20. The diameter of the bed material was used as
0.77 mm and bed roughness calculated from the bed materials using equation (1).

59
 6.2.1 Velocity Field
From the simulation, the velocity vectors at water surface were found for pre-existing
condition without barbs and with barbs, respectively. In this section, the outcomes of the flow
fields and bed-topographies are given to understand the morphodynamics resulted from
dealing of two settings: in before constructing the barbs and in the presence of the barbs.
Figure 6.1 and 6.2 show the simulation result of velocity vector without barb structure after 1
hour and 5 hours, respectively. In the figures, the simulated results of present study ( in Fig b)
compared with the previous study by Hossain et al ( 2012) (in Figure a). Figure (b) found
well agreed with Figure (a). At the time of entering the flow into the channel without barbs,
the flow towards the channel creates parabolic flow patterns that is influence to formation of
double row bar. The flow pattern is not change into entire channel after long time simulation.
Single barb has installed to the purpose of simulation and run for 1 hour and 5 hours,
respectively. Figure 6.3 and 6.4 show the simulated vector for 45º angled barb. In this
situation owing to the flow separation at barb head return current developed at the upstream
side and creates a mixing zone just behind the structure along the near bank line. The flow
passes over the structure contraction-accelerated discharge occurs at the barb end. The
convergence of these flow components result turbulent mixing around the barb head and
vector flow directed towards the outer bank near water surface.

(a)

Y (m)

x (m )
(b)
Figure 6.1 : velocity field simulation result at water surface into the channel without barb
after 1 hour simulation (a) by Hossain et al (2012) ( b ) present study (Case S3C1).

60
 (a)

x( m)
( b)

Figure 6.2 : Velocity field simulation result at water surface into the channel without barb after
5 hours ( a ) by Hossain et al ( 2012) (b) present study.

(a)

Y (m)

x(m )
(b)
Figure 6.3: Velocity field with a barb installed one side of the channel after 1 hour
simulation (a ) by Hossain et al (2012) ( b ) present study (Case S3C2).

61
 (a)

y (m)

x(m)

(b)
Figure 6.4: Velocity field with a 45° barb installed at one sidewall of the channel after 5
hour simulaion ( a ) by Hossain et. al (2012) ( b ) present study (Case S3C2).

x (m )
( a ) 1 hour

x (m )
( b ) 5 hours

Figure 6.5 : Velocity field with 90° barb installed at one side of the channel ( a ) after 1
hour simulation, (b) after 5 hours simulation (Case S2C3).

62
 x(m)
( a)

X (m )
(b)
Figure 6.6 : velocity field with 135° barb installed at one sidewall of the channel (a) after 1
hours simulation (b ) after 5 hours simulation.
Figure 6.5 shows the velocity vector with 90º barb for simulation of time 1 hour and 5 hours.
It is observed that the double row bars at upstream are not regular in the downstream of the
barbs. Figure 6.6 shows velocity field with 135º barb for simulation time of 1 hour and 5
hours. The general features such as circulation behind the barb and deflected flow at the tip of
the barb are reproduced succesfully. In this case, deflected flow at the tip of barb is found to
be weaker than other two cases. Stream barbs at both side of barb are simulated and result are
presented in figure 5.7 and 5.8. From those figure sit is observed that because of using barbs
at both side the flow component converge around the both barb head and flow vector directed
towards the centre of the channel.

63
 (a)

x(m)
(b)
Figure 6.7 : Velocity field result with double barb installed at both sides at the channel after 1
hour simulation ( a ) by Hossain et al (2016) (b) present study ( Case S3C5).

x (m )
(b)
Figure 6.8 : Velocity field result with double barb installed at both sides at the channel (b)
computed simulation after 5 hours (Case S3C5).

6.2.2 Bed Elevation Around the Structure

With sediment transport, the flow patterns were persuaded so that it can build the scour-hole
and deposition. Generally, scours and deposition phenomena depend on the flow. Fig. 6.9 and
6.10 depict the simulation result with the depositional contour lines of pre-existing channel
after 1 hour and 5 hours, respectively. In pre-existing channel double row bars are formed
because of flow. After 5 hours these formation of barbs become more perceptible. Fig. 6.11
and Fig. 6.12 represent bed level contour results of the channel with a barb after 1 hour and 5
hours simulation, respectively. In these Figures, Figure (b) is the simulated results of present
study, which is compared with the previous study by Hossain et al (2012) in Fig (a). Figure
(b) is found to be well agreed with Figure (a) in both the cases. Figure 6.13 delineates the
simulation results of a 90º barb with depositional contour lines after 1 hour and 5 hours,
respectively. Fig 6.14 illustrates the simulated contour lines of 135º barb after 1 hour and 5
hours, respectively. 6.15 and 6.16 depict the simulation result of both side barbs with
depositional contour lines after 1 hour and 5 hours, respectively. Deposition happens like
sand bars along the bank-line at the upstream side of the structures. On the upstream side the
64
 formation of bars changing trend from double row to alternate for long time simulation of the
barb field. As a result of increased potential energy in the zone of upstream through
backwater effects, the progression of the subcritical flow reaches in the bank region and
controls erosion. Ultimately, it leads to deposition of sediments along the bank-line.
Moreover, reduction of velocity gradient near bank through the subcritical backwater effect
promotes sediment deposition upstream of the barb structure. Head scour occurs due to
hydraulic jump and turbulence from flow mixing at the downstream side of barb especially
near the barb. The scour depth occurs at the barb end and downstream of the structure
because of energy redistribution away from the outer bank towards the center of the channel
results in scour near the ends of the barbs (Matsuura and Townsend, 2004). The scour result
is from contraction flow acceleration by local barb and deposited in the upward side of the
barb.

y (m)

x(m)
(b)

Figure 6.9 : Bed level contour result of the channel without any structure after 1 hour
simulation ( a) by Hossain et al (2012), ( b) present study (S3C1).

65
 (a)

y (m)

x(m)

(b)

Figure 6.10 : Bed level contour result of the channel without any structure after 5 hour
simulation ( a ) by Hossain et al (2012), ( b) present study (Case S3C1).

(a)

Y (m )

x ( m)
(b)

Figure 6.11: Bed level contour result of the channel with a barb installed at sidewall of the
channel after 1 hour simulation ( a ) by Hossain et al (2012), (b) Present study (Case S3C2).

66
 (a)

x(m)
(b)

Figure 6.12 : Bed level contour result of the channel with a barb after 5 hour simulation (a )
by Hossain et al (2012), (b) present study (Case S3C2)

x(m)
(a)

(b) x (m)

Figure 6.13: Bed level contour result of the channel with a 90° barb installed at one sidewall
of the channel ( a) after 1 hour (b) after 5 hour simulation (S3C3).

67
Y(m)

x (m)
(a)

Y(m)

x (m)

(b)

Fig 6.14: Bed level contour result of the channel with a 135° barb installed at one sidewall of
the channel ( a )after 1 hour simulation, (b) after 5 hour simulation (Case S3C4).

y (m)

x(m)

Figure 6.15: Bed level contour line after installed barb both side of the channel after 1 hour
simulation (a) by Hossain et al (2016), (b) present study (S3C5).

68
(a)

(b)
Figure 6.16: Evolution of bed profile & Comparison before installing barb (a) after 1 hour (b)
after 5 hours (Case S3C1).
Fig. 6.16 (a) and 6.16 (b) show the comparison of bed elevation after 1 hour and 5 hours,
respectively, in which present study has considerable agreement with previous numerical
results presented by Hossain et al (2012). Similarly 6.17 (a) and 6.17 (b) depict bed elevation
comparison with a 45º barb after 1 hour and 5 hours, respectively, where present study
perfectly matched with previous numerical studied by Hossain et al (2012). The result shows
clearly the applicability of barbs that reduced the erosion around the structure and increase
the pool habitat near the thalweg, which can be useful for natural rivers training work. Figure
6.18( a) and 6.18 (b) represent bed deformation for 90º angled barb after 1 hour and 5 hours,
respectively. Figure 6.19 (a) and 6.19 ( b ) show the evolution of bed profile for 135º barb.

69
Maximum scour is found for 45º angled groin which is .065 m and minimum scour is found
in case of 135º angled groin which is 0.054 m.

(a)

(b)
Figure 6.17: Evolution of bed profile and bed elevation with 45º barb (a) after 1 hour and (b)
after 5 hours.

70
(a)

(b)
Figure 6.18: Evolution of bed profile represents bed elevation with 90° barb (a) after 1 hour
(b) after 5 hours.

71
 (a)

(b)
Figure 6.19 : Evolution of bed profile bed elevation with 135° (a) after 1 hour (b) after 5
hours.

72
(a )
Figure 6.20: Evaluation of bed profile and comparison of Bed Elevation double at both side
(a) after 1 hour.

(b)
Figure 6.21: Evaluation Bed Elevation double barb at both sides of double barb (b) after 5
hours.

73
 El
ev
ati
on
(m
)

(a)
Figure 6.22 : Comparison of Bed Elevation of 45°, 90° and 135° barbs after 1 hour
simulation.
.

(b)

Figure 6.23 : Comparison of Bed Elevation of 45° , 90° and 135° barbs (b) after 5 hour
simulation.

74
 After 5 hours, equilibrium scouring is attained and the maximum scour depth for 45º, 90º and
135º are found as 0.065 m, 0.062 m and 0.054m, respectively. After 1 hours, the equilibrium
scour was not achieved. Therefore the result for maximum scour is not consistent with that of
5 hours result.

6.2.3 Flow Depth

For flow depth, comparisons were shown before barb installation and after barb installation.
Based on numerical result, it can conclude that the maximum scour depth is increased at a
decreased rate with the increasing in approaching flow depth. In case of shallow flow, scour
depth increased with flow depth proportionally. In pre-existing channel, flow depth with
scour hole varied from 2 cm to 10 cm and showed small variation in the entire simulation.
When barb installed in the river left bank, maximum depth occurs near the thalweg due to
turbulence flow. Expected flow depth is found to be occurred around the end of the barb and
bed level scouring found to be varied from 2 cm to 14 cm.

(a)

Y(m)

(b) x (m )

Figure 6.24: Water depth contour result of the channel without barb after 1 hour simulation
(a) by Hossain et. al (2012), (b ) present study (case S3C1).

75
y(m)

(a) x (m)

Y (m)

Figure 6.25: Water depth contour result of the channel with a barb installed at one sidewall of
the channel after 5 hours simulation (S3C1).

(a)

y(m)

(b) x (m)

Figure 6.26: water depth contour result of the channel with a 45° barb installed at one
sidewall of the channel after 1 hour simulation (a) by Hossain et al (2012), (b) present study
(case S3C2) .

76
y (m)

(a) X (m)

Figure 6.27: water depth contour result of the channel with a 45° barb installed at one
sidewall of the channel after 5 hours computation (case S3C2).

y (m)

x (m)
(a)

y (m)

x (m)
(b)

Figure 6.28 : water depth contour result of the channel with 90° barb installed at one sidewall
of the channel (a) after 1 hours simulation, (b) after 5 hours simulation (S3C3).

77
y (m)

x (m)

(a)

y (m)

x (m) (b)

Figure 6.29 : water depth contour result of the channel with 135° barb installed at one
sidewall of the channel after 5 hours computation (S3C4).

6.3 SIMULATION OF A CHANNEL WITH MULTIPLE SPUR DIKES

The purpose and function of spur dikes are same as barb and groin, different paper named
them differently. Experimental study presented by karami et al (2014) is reproduced by iRIC
Nays2DH to find out scouring phenomenon. They constructed a rectangular 14m long flume,
with 1m width and 1 m depth in the laboratory of the amirkabir University of Technology.
Three non-submerged, impermeable 25 cm length spur dikes were perpendicularly installed
in channel. The first dike installed at a distance of 3.16 meters from the beginning of the
channel and the distance of 3.16 meters from the beginning of the channel and the distance
between them selected twice the length.

6.3.1 Flow Domain and Hydraulic parameters

For the numerical simulation, the channel length was chosen 5 meters. The computational
mesh consist of 130 grids in longitudinal direction and 26 grids in lateral direction. The
constant discharge for S4C1 is 0.035 m3/s and S4C2 is 0.046 m3/s. The depth of water is kept

78
constant at 15 cm. Manning’s roughness co-efficient is considered 0.02 in this simulation.
The simulation are performed in moveable bed condition with bottom slope, S0 = 0.001. The
finite differential methods of advection term are used as CIP method.

Pourshahbaz et al (2017) also used same experimental data of Karami et al (2014) to validate
their model. In their model, they used k-ε turbulence model with the development of
Renormalized group (RNG). According to them, changing the channel dimensions seem
more reasonable and the length of channel should not be too short or too long to minimize
computing costs.

Table 6.2 Hydraulic parameter for this numerical simulation

No. of L B Y Q U Fr d50
test
m m m m3/s m/s - mm
S4C1 5 1 0.15 0.035 0.233 0.19 0.91
S4C2 5 1 0.15 0.046 0.307 0.25 0.91

Here, L = Length of the channel, B=width, Y= depth of water, Q=Discharge, U= velocity, Fr

= Froude number and D50 = Sediment size. S4C1 and S4C2 represents two different test with
different hydraulic parameter.

6.3.2 Numerical Simulation Result

Numerical simulation by iRIC Nays2DH are performed for three perpendicular spur dikes
with two different discharge. According to the result which are shown in Figures 6.30 and
6.31, maximum scour is 0.145m for Case S4C1 and 0.24 m for case S4C2. Figure 6.32 and
6.33 show the computed velocity vector for case S4C1 and S4C2, respectively. It is observed
that spur dikes influenced the flow at downstream zone and created a mixing zone among the
spur dikes. Figure 6.34 shows the definition sketch for different cross and longitudinal
sections. Two longitudinal and two cross-sections are taken for whom the sectional bed
elevation are compared in the following section.

Figure 6.35 shows the bed level changes comparison for case S4C1 with previous
experimental results. Maximum scouring are found in Figure 6.35 ( d). For case S4C1, the
scour depth from first spur dike to third spur dikes are found as 0.123m, 0.005m and 0.005m,
respectively. For case S4C2, the scour depth from first spur dike to third spur dikes are found
as 0.16 m, 0.01m and 0.012m, respectively (Figure 6.36 (d)). From simulated results, it is
observed that maximum scouring occurs at the first spur dike. The amount of scouring and
sedimentation are also found to be depended on flow discharge. From the comparison of
results between two cases, it can be concluded that higher discharge is responsible for higher
scouring and sediment deposition. Table 6.3 shows the comparison of the scouring amount
found by present study with previous studies of FLOW-3D simulation (pourshahbaz et al.,

79
2017) and previous experimental results (Karami et al., 2014). The results are found to be
consistent with each other.

Figure 6.30 : Bed deformation contour line of case S4C1 simulated by iRIC Nays2DH

Figure 6.31 : Bed deformation contour line of case S4C2 simulated by iRIC Nays2DH.

80
 Figure 6.32: Computed velocity vector in case of case S4C1 developed by iRIC Nays2DH.

Figure 6.33 : Computed Velocity vector in case of S4C2 developed by iRIC Nays2DH.

a b

d
d

0.35m c
c 0.15m
0.25m

3.16m a b
3.41m
Figure 6.34 : Definition sketch for different cross and longitudinal sections.

81
 (a) Transverse bed profile at X= 3.16 ( section a-a)

present study

experiment( Karami et al., 2014)

0.04 FLOW3D ( Pourshahbaz et al.,2017)
0.02
0
-0.02
-0.04
z(m)

-0.06
-0.08
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

y (m)

( b) Transverse bed profile at x =3.41(section b-b)

82
 present study

experiment ( karami et al.,2014)

0.04 FLOW3D (Pourshabaz et al.,2017)

0.02
0
-0.02
-0.04
Z (m)

-0.06
-0.08
-0.1
-0.12
-0.14
2.8 3 3.2 3.4 3.6 3.8 4

x (m)

(c ) Longitudinal bed profile at y = 0.15 (section c-c)

(d) Longitudinal bed profile at y = 0.35 (section d-d)

Figure 6.35: Bed changes comparison of case S4C1 results found by iRIC Nays2DH with
FLOW-3D and previous experimental results.

83
 (a) Transverse bed profile at X = 3.16 (section a-a)

present study
experiment ( Karami et al., 2014)
FLOW3D (Pourshahbaz et al.,2017)
0.1
0.05
0
Z (m)

-0.05
-0.1
-0.15
-0.2
-0.25
0 0.2 0.4 0.6 0.8 1
y (m)

(b) Transverse bed profile at X = 3.41(section b-b)

84
( c ) Longitudinal bed profile at y = 0.15 (section c-c)

(d ) Longitudinal bed profile at y = 0.35 (section d-d)

Figure 6.36 : Bed changes comparison of case S4C2 results found by iRIC Nays2DH with
FLOW-3D and experimental results.

85
Table 6.3 : Comparison of the scouring amount in iRIC Nays2DH results with Flow-3D
(Pourshahbaz et al., 2017) and experimental results (Karami et al., 2014).

Case ds1 ds2 ds3

Experimental 0.156 0 0.026

S4C1 FLOW-3D 0.123 0.005 0.023
iRIC Nays2DH 0.123 0.005 0.005

Experimental 0.225 0.029 0.072

S4C2 Flow-3D 0.178 0.0177 0.0321
iRIC Nays2DH 0.16 0.01 0.012

86
 Chapter 7

CONCLUSION AND RECOMENDATION

7.1 INTRODUCTION

The study employed two dimensional simulation model Nays2DH to investigate the
characteristics of flow and bed change around a groin like structures. The simulation is
performed for single and multiple groins with different discharges. To examine the
performance of software with strong secondary current, open channel flow with 180º sharply
and mildy curved channels are also simulated.

7. 2 Conclusions

For all the channel bends, the simulation predicts the flow field accurately at the inlet straight
portion, at bend as well as at the outlet straight portion of the channel. The simulated flow
field and water surface profile for 1800 bends obtained from the two-dimensional model
showed good comparison with experiments. Therefore, iRIC Nays2D is suitable to deal a
flow field which is highly affected by secondary current.
This simulation has provided us with detailed information regarding the flow pattern, the
velocity and bed shear stress profiles around a groin. This adds to the effort of work done by
others on different types of groins. The flow features around a groin is reproduce
successfully. The flow field is found to be divided into four distinct regions: (i) the uniform
flow at upstream end, (ii) big circulation of low velocity zone, (iii) a small low velocity zone
near the foot at the upstream of groin and (iv) high velocity zone opposite to the circulation
zone created due to the deflected flow. The high velocity zone is found to be decayed toward
downstream and the low velocity in circulating zone is found to be regained. The computed
velocity and bed shear stress profiles of presents study are compared among different cases.
From the simulation, it is observed that the peak of velocity and bed shear stress is found
maximum at the position of head of groin when lateral distance y/l =1, where l is the groin
length. The position of maximum velocity and bed shear stress is found to be shifted towards
downstream with increasing y/l. The maximum velocity and bed shear stress for 135º groin
are found are found lower than the other two cases for all the section of y/l. W/U0 and /
for 90º groin are found maximum for y/l ≤ 2, and for y/l ≥ 2, the velocity and bed shear stress
are maximum for 45º angled groin. The downstream circulation zone is found to be larger for
45º groin compared to 90º groin; similarly, the circulation zone for 90º groin is found larger
compared to 135º groin. In order to examine the present model, the computed results are
compared with previous numerical and experimental results. The comparison are found to be
well agreed with each other.

87
A barb is installed with 45º angled to the flow direction similar to Hossain et al (2012) in a
straight channel and simulated for 1 hour and 5 hours. Same hydraulic data is also used for
90º as well as 135º angled barb and also simulated for 5 hours. After 5 hours simulation, the
flow is found to be reached in equilibrium scouring. In that case maximum scouring has
found for 45º angled groin and minimum scouring has found for 135º angled groin. After
barb installation, due to the flow separation at barb head, return current developed at the
upstream side induced zone of subcritical flow along the stream bank. It influenced the flow
at downstream side creates a mixing zone just behind the structure. At the downstream side of
barb especially near the barb head scour occurred due to hydraulic jump and turbulence from
flow mixing. In pre-existing channel flow depth with scour hole varied from 1 cm to 8cm. As
expected flow depth was present around the end of the barb and the water depth including
scour hole varied between 2 cm to 12 cm. Simulation results showed that the installed barb
successfully reduced the flow velocity along the bank.

The numerical simulation is also performed for three perpendicular groin in a straight
channel. Two different test S4C1 and S4C2 were run for 8 hours with 0.035 m3/s and 0.046
m3/s, respectively. Maximum scouring occurred at first spur dike than the other two spur
dikes. Considerable agreements are found between computed and previous experimental
results. It is found that the amount of scouring and sedimentation are also depend on flow
discharge. From the simulation results, it can be concluded that higher discharge is
responsible for higher scouring and sediment deposition.

7.2 RECOMMENDATIONS FOR FURTHER STUDY

Following recommendation can be made for further research on this topic.
 Since the flow near a groin is highly three-dimentional, it is recommended that for
field, a 3D numerical model needs to be employed.
 The study can be extended to simulate the sediment and nutrient transport near a groin
using 3D software.
 The model can be applied for different spacing between two groins to study the effect
of spacing.
 The study can be extended for different groin length to channel width ratios to study
the effect of groin length to the flow field and sediment transport.

88
 REFERENCES
Ahmad, M. ( 1951), Spacing and protection of spurs for bank protection, Civil Engineering
and Publication Review, 46, 3-7.

Ahmad, M. (1953), Experiments on design and behaviour of spur dikes. In Proceedings:

Minnesota International Hydraulic convention (pp. 145-159). ASCE.

Barbhuiya, A.K., & Dey, S. (2004). Local scour at abutments: A review. Sadhana, 29 (5),
449-476.

Chow, V.T. (1959). “Open-channel hydraulics”. Mc Graw-Hill Book co., Inc., New York,
N.Y.

De-Vriend, H.( 1980).“Velocity redistribution in curved rectangular channels.” J. Fluid

Mech., 107, 423-439.

De-Vriend, H. (1981). “Flow measurement in a curved rectangular channel. Part 2: Rough

bottom.” Internal Rep. No. 5-81. Tech. Rep., Laboratory of Fluid Mechanics, Dept. of Civil
Engineering, Delft Univ. of Technology, Delft, The Netherlands.

Davinroy, R. D.(1990). Bendway Weir Concept in the Mississippi River, Proceedings,

International Association of Navigation Congress, Secille, Spain, pp. 115-123.

Davinory, R. D. (1994). The Bendway Weir, A New Structural Solution to Navigation

Problems Experienced on the Mississippi River, International Association of Navigation
Congresses, Bulletin 69, Osaka, Japan, pp. 5-19.

Fox, j.F., Papanicolaou, A.N., and Kjos, L.(2002). An Environmentally Friendly Barb design
for Washington state, Research and Extension Regional Water Quality Conference, February
20-21, Vancouver, WA.

Garde, R., Subramanya, K., & Nambudripad, K.D. (1961), Study of scour around spurdikes,
Journal of Hydraulics Division, 86, 23-37.

Gill, M.A. (1972), Erosion of sand beds around spur dikes. Journal of Hydraulic Divison, 98

Hossain, S.A.A.M, Uddin, M.J., Ali, M. S. (2012). “Effect of stream barb on straight channel
bed configuration: Numerical simulation”. ICCESD-2012, KUET, Khulna, Bangladesh.

Hasen, E. M. and Winter, E. R. (1996). Abstract 0027: design abstract for barb. Northwest
Region Hydraulic Section (Draft).

89
 Jamieson, E.C., Rennie, CD., Townsend, R.D. (2013). Turbulence and Vorticity in a
Laboratory Channel Bend at Equilibrium Clear-Water Scour with and without Stream Barbs.
J. Hydraul. Eng., ASCE, 139: 259-268.

Kim, T.B. and Choi, S.U. (2003). “conducted an experiment on numerical simulations of
open-channel flows in a bend using the finite element method”. School of Civil and
Environmental Engineering, Yonsei University, Seoul, Korea.
.
Kuhnle, R. A., Alonso, C.V., & Shields, F.D. (1999), The geometry of Scour holes associated
with 90 spur dikes, Journal of Hydraulic Engineering, 125 (9), 972-978.

Kothyari, U.C., & Rangaraju, K.G. (2001), K.G. (2001), Scour around spur dikes and bridge
abutments. Journal of hydraulic research, 39 (4), 367-374.

Karami, H., Basser, H., Ardeshir, A., & Hosseini, S.H. (2014), Verification of numerical
study of scour around spur dikes using experimental data. Water and Environment journal,
28(1), 124-134.

Kang, J., Yeo, H., Kim, S., & Ji, U. (2011). Experimental investigation on the local scours
characteristics around groynes using a hydraulic model. Water and Environmental Journal,
25 (2), 181-191.

Lloyd, P.M. and Stansby, P. K. ( 1997). Shallow-water flow around model conical islands of
small side ΙΙ: submerged. J. Hyd. Engr., Asce, 123 (12), pp. 1068-1077.

NRCS (2005), Technical Note 23, Design of stream Barbs (Version 2.0). U.S. Department of
Agriculture, Portland, Oregon. Oregon Bulletin No. OR210-2005, pp 1-25.

Quanhong, L., Pengzhi, L.(2007). Numerical Simulation of Recirculating Flow Near a

Groyne. The 2nd International Conference on Marine Research and Transportation. Ischia.
Naples. Italy, pp. 61-68.

Rajaratnam, N. and Nwachukwu, B.(1983) ‘Flow near Groin-like Structure’,Journal of

Hydraulic Engineering, 109 (3): 463-480.

Rozovskii, I.L. (1961). “Flow of water in bends of open channels.” Academy of Science of
krainian SSR, Russia.

Shimizu, Y. (2002). A method for simultaneous computation of bed and bank deformation of
a river, Proceedings of the International Conference on Fluvial Hydraulics, River Flow 2002,
Louvain-la-Neuve, Belgium, vol. 2, pp 793-802, September, 2002.

90
Schiechtl, H. M. and R. Stern. (1996). Water Bioengineering Techniques for Watercourse,
Bank and Shoreline protection. Trans by L. jaklitsch. Blackwell Scientific. Oxford, UK.
P.185.

Sarveram, H., Shamsai, A., Banihashemi, M.A. (2012). Two-dimensional simulation of flow
pattern around a groyne using semi-implicit semi-Lagrangian method. International Journal
of Physical Science, Academic Journals, 7(20): 2775-2783.

Shields Jr, F.D., Cooper, C.M., & Knight, S.S. (1995), Experiment in stream restoration
Journal of Hydraulic Engineering, 121(6), 494-502.

Tingsanchali, T. and Maheswaran, S.(1990) ‘2-D Depth-averaged Flow Computation near

Groyne’ Journal of Hydraulic Engineering, 116(1):71-86.

USDA, (2001). Design of Stream Barbs, Technical Notes, Engineering No. 12, Version 1.4,
US Department of Agriculture, Natural Resources Conservation Service, Boise, ID.

91